Computational Intelligence

This paper introduces a Genetic Algorithm (GA) based optimization for rule base and scaling factors adjustment to enhance the performance of fuzzy logic controllers. First a recursive rule base adjustment algorithm is developed, which has the benefit that it is computationally more efficient for the generation of a decision table. Then utilizing the advantage of GA optimization, a novel approach that each random combination of the optimized parameters (including the membership function selection for the rule base and controller scaling factors) is coded into a Real Coded string and treated as a chromosome in genetic algorithms is given. The optimization for rule base with the correspondent membership function and scaling factors using GA is easy to be realization in engineering. Simulation results are presented to support this thesis.

This paper considers a generalized fuzzy version of maximum flow problem in which arc capacities are fuzzy variables. The problem is to find a maximum flow under some chance constraints with respect to credibility measure of arc flow of network. Some crisp equivalents of fuzzy chance constraints are presented when the fuzzy capacities are characterized by trapezoidal fuzzy numbers and general fuzzy numbers respectively. Furthermore, a genetic algorithm is used to solve the maximum flow with these crisp equivalents. Finally, a numerical example is provided for the effectiveness of the model and algorithm.

This paper concerns a problem of delay-dependent stability of a class of nonlinear continuous-time systems with time delays, based on the fuzzy hyperbolic model (FHM). FHM is a universal approximator, and can be used to establish the model for unknown complex systems. Moreover, the main advantage of using FHM over T-S fuzzy model is that no premise structure identification is need and no completeness design of premise variables space is need. Also an FHM is a kind of valid global description and nonlinear model inherently. The proposed method is addressed by solving a set of linear matrix inequalities (LMIs). A general Lyapunov-Krasovskii functional are used and some additional free-weighting matrices, which bring much flexibility in solving LMIs, are introduced during the proof. The results then have easily been extended to a system with either polytopic-type or norm-bounded uncertainties. A simulation example is given to validate the proposed results.

Water quality management plays a very important role in fish life and aquatic product quality. The paper firstly selects the index system and confirm the weight of each of the water quality factors dependent on the fish tolerance to the factors of water quality integrated with the result of the expert questionnaire by the Delphi method with the Water Quality Standard, then constructs the fuzzy evaluation model of the multiplex water quality parameter and classifies the standard of water quality into five classes of standard. At last the paper give an experimental model based on the monitored datum in North China. It shows that it can evaluate the water quality integrative and provide the degree of membership that the water quality belongs to all the standards.

To study the joint velocity vector trajectory regulation and control method of a redundant space manipulator. A novel multi-restriction manipulator joint velocity vector control algorithm, based on fuzzy reasoning theory, is presented. The task executed by the redundant manipulator is broken into a series of sub-tasks expressed with vectors. The subordinate task is executed in the redundancy space of executing the primary task. The conventional joint velocity vector algorithm is combined with fuzzy reasoning theory so that every subordinate task is best optimized. Singular gesture and arithmetic singularity are avoided. The multi-restriction redundant manipulator joint velocity vector control problem is solved. The algorithm validity is proved by the numerical simulation results.

This paper develops a fuzzy hyperbolic control method for chaotic continuous-time systems with uncertainties. First, the generalized fuzzy hyperbolic model (GFHM) is used to model unknown part of a chaotic system. Second, based on Lyapunov functional approach, a sufficient condition for a fuzzy hyperbolic controller and a state feedback controller is given such that the closed-loop system is asymptotic stable. Moreover, considering the influence of both approximation error and external disturbance, fuzzy hyperbolic H_∞ control scheme is addressed . All the results are given in terms of LMI forms, the effectiveness of the proposed method is shown by a simulation example.

In context-aware systems, one of the main challenges is how to tackle context uncertainty well, since perceived context always yields uncertainty and ambiguity with consequential effect on the performance of context-aware systems. We argue that uncertainty is mainly generated by two sources. One is sensor’s inherent inaccuracy and unreliability. The other source is deduction process from low-level context to high-level context. Decision tree is an appropriate candidate for reasoning. Its distinct merit is that once a decision tree has been constructed, it is simple to convert it into a set of human-understandable rules. So human can easily improve these rules. However, one inherent disadvantage of decision tree is that the use of crisp points makes the decision trees sensitive to noise. To overcome this problem, we propose an alternative method, fuzzy decision tree, based on fuzzy set theory.

Based on integrating the property of sliding mode control (SMC) with the thought of variable universe in adaptive fuzzy control, a design method of variable universe adaptive fuzzy sliding mode control (FSMC) is proposed. There are two sets of control rule bases. The first set is utilized to approach the equivalent control of SMC. By adjusting the universes of input variables and the membership fuzzy controller of conclusion part in rules on-line, a variable universe adaptive fuzzy control is developed to estimate the equivalent control of SMC control system. The derived adaptive law is applied to adjust the rule parameter for changing the control rules to meet system dynamic. Another set is used to attenuate the switching control of SMC in the sense of heuristic, which ensure the requirement of system stability. Four heuristic control rules are employed to smooth the control law based on the concepts of SMC. We apply the control method to the missile electro-hydraulic servo mechanism. Simulation results verify the validity of the proposed approach.

This paper presents a simple and useful decision method for making estimates in uncertain and dynamic environments. We extend traditionally triangular fuzzy numbers to pillared and triangular fuzzy numbers formats that are defined by X-Y-Z axes. The α-cut and the defuzzy method of the proposed fuzzy number type are also defined. Finally, a numerical example based on a multiple attribute decision making method is used to illustrate the proposed method.

A novel Fuzzy PID controller is proposed in this paper, which is used for treating the control of pulp level and brightness in flotation de-inking column. The essential part of fuzzy PID controller is the fuzzy logic control, which is a multi-input-multi-output controller. A design method of the fuzzy PID controller is presented. The simulation results show that the proposed Fuzzy PID controller has preferable performance and significant advantages over the traditional PID controller.

In universal logic, S-norm is the mathematical model of “OR” operation. S-norm and S-generator were defined on interval [0, 1] in previous work. In the related work, authors put forward a kind of logic based on generalized interval [a, b]. This paper studied the S-norm and S-generator on any interval [a, b], discussed the two kinds of generalized S-generators: “Automorphic increase S-generator” and “Infinite decrease S-generator”, and proved the important generating theorem of generalized S-norm. Based on the conceptions of integrated cluster of generalized S-norm and S-generator, authors put forward a new uniform “OR” operation model. The simulation shows that it is not only flexible but controllable. It enlarged the study domain of universal logic, and offered important theory for uncertainties reasoning of complex system.

It is necessary to find the relationship between product information and development process for saving product development time. Design information constraint tree (DICT) is proposed to describe the constraint relationships among multi-task information. Fuzzy timed workflow net (FTWN) is introduced to analyze product development time. An algorithm is proposed to map DICT onto FTWN. Task execution time and completion possibility are analyzed. The effectiveness of the proposed method to simulate product development time is validated by an example.

Considering the indefinite and fuzzy characters existing in index weighs of complex system, a method confirming index weights based on evidence theory was put forward. On the basis of it, an efficiency evaluation model for complex system integrating Analytical Hierarchy Process (AHP) with fuzzy synthetic evaluation method has been set up, which can obtain more reasonable and objective evaluation result for complex system. Through analyzing an example of the system fighting efficiency for missile and gun integration weapon, the result shows the validity and feasibility of this method.

In this paper, several kinds of non-classical fuzzy problems–direction fuzzy problems are presented, which classical fuzzy sets do not deal with well. In order to solve them, direction fuzzy sets are introduced by breaking through the limitations of the definitions of classical fuzzy sets. And the union and intersection of direction fuzzy sets are discussed. Especially, an application example (Escape Model) is dug up to show the significance of the direction fuzzy set theory.

This paper presents a new approach using fuzzy compensator and PID controller to an experimental propeller setup which is called the twin rotor multi-input multi-output system (TRMS). Some previous works ignored the interactions between two axes and the controller being designed in horizontal or vertical direction separately. The goal of this study is to stabilize the TRMS in significant cross coupling conditions and to experiment with trajectory tracking. The fuzzy compensator and PID controller design is performed by a real-valued GA (RGA) with system performance index as fitness function. We apply the integral of time multiplied by the square error criterion to form a suitable fitness function in the RGA. Simulation results show that the proposed design can successfully adapt system nonlinearity and complex coupling conditions.

Fuzzy neural network adaptive tracking controller is designed to realize the tracking control for a class of unknown nonlinear interconnected systems. No constraint or matching conditions of the uncertain terms are required. For the low dimensions unknown dynamic of the subsystems and the high one of the interconnected terms, two classes of fuzzy rules are adopted respectively to approximate the unknowns. The neural network is used to counteract the extra gains of the controller. The fuzzy sliding mode control is developed to compensate for the exterior disturb and the fuzzy neural network approximation errors. By the Simultaneity, based on Lyapunov method, the parameters of the systems are regulated on line by the adaptive laws. Global asymptotic stability is assured with the tracking errors converging to a neighborhood of zero.

To measure the certainty, we use the meaning of entropy. For the selection of reliable data, fuzzy entropy through distance measure is proposed. The appropriateness of the proposed entropy is verified by the definition of entropy measure. To measure the fuzziness of 3-phase stator currents, membership functions are obtained by the Bootstrap method. Finally, the proposed entropy is applied to the membership function of 3-phase currents, and the fuzzy entropy values of phase current each are illustrated.

The Fuzzy Inverse Kinematic Mapping Method (FIKM) is used to solve the inverse kinematics for the redundant robots with singularities. This method has some advantage due to the less computation load and robustness to the singularity. The method has also been implemented on a transputer-based parallel processing system to solve the motion planning problem of the redundant robots with singularity.

This paper presents a design method of robust passive controllers for continuous-time Takagi-Sugenon (T-S) fuzzy systems. The parametric uncertainty is assumed to be norm bounded. By applying the Lyapunov stability theory, a sufficient condition on the existence of robust passive controllers is given. With the help of linear matrix inequality (LMI), robust passive controllers are designed such that the closed-loop system is robust stable and satisfies the passive performance. Furthermore, a convex optimization problem with LMI constraints is formulated to design robust passive controller with the maximum dissipation rate. A numerical example demonstrates the validity of this method.

A robust controller design method based on the concept of fuzzy region is presented for Takagi-Sugeno (T-S) fuzzy systems with parametric uncertainties. The uncertain T-S fuzzy model is converted into uncertain T-S fuzzy region model. A relaxed sufficient condition is derived for robust stabilization by using Lyapunov function approach in the form of linear matrix inequalities (LMIs). Comparing with other similar conditions, the difficulty of solving LMIs is greatly reduced in this method. The efficiency of the method is illustrated through the simulation of a numerical uncertain fuzzy system.

In practice, input and rule perturbation are two important factors that will heavily influence the performance of fuzzy system. Quantitatively measure the influence of these two kinds of perturbation on the input/output mapping relationship of fuzzy system has great significance, theoretically and practically. In this paper, a statistical-based quantitative measurement for input and rule perturbation is proposed. By using the proposed approach, influence of perturbations on fuzzy system can be computed quantitatively, analytically and efficiently. Simulation results demonstrate the effectiveness of the proposed approach.

Application of agent technology in Intrusion Detection Systems (IDSs) has been developed. Intrusion Detection (ID) agent technology can bring IDS flexibility and enhanced distributed detection capability. The security of the ID agent and methods of collaboration among ID agents are important problems noted by many researchers. This paper applies fuzzy logic to reduce the false positives that represent one of the core problems of IDS. ID is a complicated decision-making process, generally involving enormous factors regarding the monitored system. A fuzzy logic evaluation component, which represents a decision agent model of in distributed IDSs, considers various factors based on fuzzy logic when an intrusion behavior is analyzed. The performance obtained from the coordination of an ID agent with fuzzy logic is compared with the corresponding non-fuzzy type ID agent.

To efficiently recover the lost motion vectors of corrupt macro- blocks, we introduce a new fuzzy metric based on Sugeno fuzzy integral as distortion metric in temporal error concealment. The proposed fuzzy metric can suit the HVS (human visual system) fairly well. Also, we integrate the proposed algorithm into H.264/AVC decoder. Experimental results from decoding several types of video sequences have shown that our scheme can considerably improve the visual quality of reconstructed images and achieve the PSNR gain at different frame.

The universal approximation property of binary-tree hierarchical fuzzy systems (HFS) is examined in this paper. A binary-tree hierarchical fuzzy system with typical FLUs (fuzzy logic units) is defined. The analytical expression of HFS is derived, and the system is shown to preserve universal approximation property. A simple example is also given to show the theory.

Local model interpretability is a very important issue in neurofuzzy local linear models applied to nonlinear state estimation, process modelling and control. This paper proposes a new fuzzy membership function with desirable properties for improving the interpretability of neurofuzzy models. A learning algorithm for constructing neurofuzzy models based on this new membership function and a hybrid objective function is derived as well, which aims to achieve optimal balance between global model accuracy and local model interpretability. Experimental results have shown that the proposed approach is simple and effective in improving the interpretability of Takagi-Sugeno fuzzy models while preserving the model accuracy at a satisfactory level.

This paper presents a novel fuzzy clustering method named as AINFCM, which solves the traditional fuzzy clustering problems by searching for the optimal centroids of clusters using artificial immune network technology. Based on the clone and affinity mutation principals of biological immunity mechanism, containing memory cells, the AINFCM is capability of maintaining local optima solutions and exploring the global optima defined as minimum of the objective function. The algorithm is described theoretically and compared with classical K-means, K-medoid, FCM and GK Clustering methods using PC, CE, SC, S, ADI and DI validity indexes. Parameter setting was also discussed to analyze how sensitive the AINFCM is to user-defined parameters.

This paper presents a new Fuzzy Neural Network (FNN) model to evaluate design alternatives in conceptual design. In the proposed method, a fuzzy reasoning based on feedforward neural network is used to evaluate concepts, and a learning algorithm based on ranking-based adaptive evolutionary operator genetic algorithm (RAOGA) is utilized to adjust fuzzy weights and thresholds with fuzzy inputs and outputs in FNN.

This study modifies the Evolving Fuzzy Neural Network Framework (EFuNN framework) proposed by Kasabov (1998) and adopts a weighted factor to calculate the importance of each factor among these different rules. In addition, an exponential transfer function (exp (-D)) is employed to transfer the distance of any two factors into the value of similarity among different rules, thus a different rule clustering method is developed accordingly. The intensive experimental results show that the WEFuNN performs very well when applied in the PCB sales forecasting.

A novel fuzzy neural classifier and learning algorithm are proposed based on EM learning in this paper. The method firstly applies rough set of its information measurement ability to evaluate system parameters importance. Then, based on EM learning the unknown parameters of fuzzy member functions are estimated. Then a fuzzy neural classifier based on EM algorithm is generated. The research indicates that the proposed network possesses higher diagnosis precision and speed as well as excellent anti-interference abilities, and is an ideal pattern classifier. In the end, a practical application in transformer fault diagnosis shows the availability of the method.

The conventional fuzzy CMAC can be viewed as a basis function network with supervised learning, and performs well in terms of its fast learning speed and local generalization capability for approximating nonlinear functions. However,it requires an enormous memory and the dimension increase exponentially with the input number. Hierarchical fuzzy CMAC (HFCMAC) can use less memory to model nonlinear system with high accuracy. But the structure is very complex, the normal training for hierarchical fuzzy CMAC is difficult to realize. In this paper a new learning scheme is employed to HFCMAC. A time-varying learning rate assures the learning algorithm is stable. The calculation of the learning rate does not need any prior information such as estimation of the modeling error bounds. The new algorithms are very simple, we can even train each sub-block of the hierarchical fuzzy neural networks independently.

While the Resilient Backpropagation (RPROP) method can be extremely fast in converging to a solution, it suffers from the local minima problem. In this paper, a fast and reliable learning algorithm for multi-layer artificial neural networks is proposed. The learning model has two phases: the RPROP phase and the gradient ascent phase. The repetition of two phases can help the network get out of local minima. The proposed algorithm is tested on some benchmark problems. For all the above problems, the systems are shown to be capable of escaping from the local minima and converge faster than the Backpropagation with momentum algorithm and the simulated annealing techniques.

The manifold learning methods can discover the varying intrinsic features in face image space. However, in order to efficiently solve face image recognition problem with an image database, the extraction of discriminative features should be firstly considered. This paper proposes a new discriminative manifold learning method for face recognition. Besides like the recently proposed local perserving projectioin and local discriminative embedding algorithms which can preserve the local structure similarity in the face submanifold, our method emphasizes the discriminative property of embedding much more by a proposed Fisher Manifold Discriminant Embedding (Fisher MDE) criterion to build an object function and achieve the maximum. Experimental results on three open face datasets indicate the proposed method achieves lower error rates and provides a promising performance.

This paper presents a new parametric learning scheme, namely, Reformulated Parametric Learning (RPL). Instead of learning the parameters directly on the original model, this scheme reformulates the model into a simpler yet equivalent one, and all parameters are estimated on the reformulated model. While a set of simpler equivalent models can be obtained from deriving Equivalent Decomposition Models (EDM) through their associated ordinary differential equations, to achieve the simplest EDM is a combination optimization problem. For a preliminary study, we apply the RPL to a simple class of models, named ’Additive Pseudo-Exponential Models’ (APEM). While conventional approaches have to adopt nonlinear programming to learn APEM, the proposed RPL can obtain equivalent solutions through Linear Least -Square (LLS) method. Numeric work confirms the better performance of the proposed scheme in comparing with conventional learning scheme.

Feature extraction plays an important part in pattern recognition (PR), data mining, machine learning et al. In this paper, a novel supervised feature extraction algorithm based on continuous divergence criterion (CDC) is set up. Firstly, the concept of the CDC is given, and some properties of the CDC are studied, and proved that CDC here is a kind of distance measure, i.e. it satisfies three requests of distance axiomatization, which can be used to measure the difference degree of a two-class problem. Secondly, based on CDC, the basic principle of supervised feature extraction are studied, a new concept of accumulated information rate (AIR) is given, which can be used to measure the degree of feature compression for two-class, and a new supervised feature extraction algorithm is constructed. At last, the experimental results demonstrate that the algorithm here is valid and reliable.

One of the powerful classifiers is Support Vector Machine (SVM), which has been successfully applied to many fields. Despite its remarkable achievement, SVM is time-consuming in many situations where the data distribution is unknown, causing it to spend much time on selecting a suitable kernel and setting parameters. Previous studies proposed understanding the data distribution before classification would assist the classification. In this paper, we exquisitely combined with clustering and classification to develop a novel classifier, Clustering-Launched Classification (CLC), which only needs one parameter. CLC employs clustering to group data to characterize the features of the data and then adopts the one-against-the-rest and nearest-neighbor to find the support vectors. In our experiments, CLC is compared with two well-known SVM tools: LIBSVM and mySVM. The accuracy of CLC is comparable to LIBSVM and mySVM. Furthermore, CLC is insensitive to parameter, while the SVM is sensitive, showing CLC is easier to use.

In the field of cluster analysis and data mining, fuzzy c-means algorithm is one of effective methods, which has widely used in unsupervised pattern classification. However, the above algorithm assumes that each feature of the samples plays a uniform contribution for cluster analysis. To consider the different contribution of each dimensional feature of the given samples to be classified, this paper presents a novel fuzzy c-means clustering algorithm based on feature weighted, in which the Variable Precision Rough-Fuzzy Sets is used to assign the weights to each feature. Due to the advantages of Rough Sets for feature reduction, we can obtain the better results than the traditional one, which enriches the theory of FCM-type algorithms. Then, we apply the proposed method into video data to detect shot boundary in video indexing and browsing. The test experiment with UCI data and the video data from CCTV demonstrate the effectiveness of the novel algorithm.

Applying Bayesian approach to decision tree (DT) model, and then a Bayesian-inference-based decision tree (BDT) model is proposed. For BDT we assign prior to the model parameters. Together with observed samples, prior are converted to posterior through Bayesian inference. When making inference we resort to simulation methods using reversible jump Markov chain Monte Carlo (RJMCMC) since the dimension of posterior distribution is varying. Compared with DT, BDT enjoys the following three advantages. Firstly, the model’s learning procedure is implemented with sampling instead of a series of splitting and pruning operations. Secondly, the model provides output that gives insight into different tree structures and recursive partition of the decision space, resulting in better classification accuracy. And thirdly, the model can indicate confidence that the sample belongs to a particular class in classification. The experiments on music style classification demonstrate the efficiency of BDT.

In this paper, we study clustering with respect to the k-modes objective function, a natural formulation of clustering for categorical data. One of the main contributions of this paper is to establish the connection between k-modes and k-median, i.e., the optimum of k-median is at most the twice the optimum of k-modes for the same categorical data clustering problem. Based on this observation, we derive a deterministic algorithm that achieves an approximation factor of 2. Furthermore, we prove that the distance measure in k-modes defines a metric. Hence, we are able to extend existing approximation algorithms for metric k-median to k-modes. Empirical results verify the superiority of our method.

Decomposition methods is the main way for solving support vector machines (SVMs) with large data sets. In this paper a new decomposition algorithm is proposed, and its convergence is also proved.

Learning Bayesian network is a problem to obtain a network that is the most appropriate to training dataset based on the evaluation measures given. It is studied to decrease time and effort for designing Bayesian networks. In this paper, we propose a novel online learning method of Bayesian network parameters. It provides high flexibility through learning from incomplete data and provides high adaptability on environments through online learning. We have confirmed the performance of the proposed method through the comparison with Voting EM algorithm, which is an online parameter learning method proposed by Cohen, et al.

Quality of Service (QoS) and Traffic Engineering (TE) capabilities are two important techniques in IP networks to support real-time applications. Multi Protocol Label Switching (MPLS) plays an important role in terms of QoS and TE provisioning. The paper investigates the optimal resource allocation in MPLS-based VPN. Firstly the hose model is introduced; then, the traffic engineering using hose model to provide multi-VPN services in MPLS network is formulated. A genetic algorithm (GA) for the optimization of the bandwidth assignment in hose-modeled VPN to provide multi services is presented. The optimal model and instances of the algorithm are given too. The results of the instances show that the proposed method can balance the network load and is simple to be implemented.

Agent-based production scheduling is a promising approach to solve production scheduling problem, especially in a dynamic, uncertain environment. In the system, agents are connected through a network and negotiate with each other to fulfill scheduling. The negotiation mechanism specifies the way in which negotiation should take place. This paper proposes an adaptive negotiation framework and two kinds of negotiation policies to fulfill scheduling and rescheduling in the flexible job shop. The mechanism makes the system more adaptive in dynamic production environments. The computational experiments are given to demonstrate the feasibility and performance of the mechanism.

The speed and convenience of the Internet facilitated the development of electronic commerce (e-commerce). E-commerce research and technologies have always drawn the attention of researchers. Among them, the application of mobile agent on e-commerce has drawn much attention in recent years. Mobile agents can roam freely over different execution environments to execute tasks assigned to them. However, a mobile agent may be attacked when it requests services from other servers or when comes in contact with and exchange information with another agents while roaming on the internet. Hence, a mobile agent user may be concerned that his mobile agent could be corrupted or private information tapped and pirated by other agents. To ensure the security of mobile agents in public network environment, this paper proposes a security scheme that is suitable for mobile agents. The scheme includes access control and key management; it is also an improvement on the key management and access control for mobile agent scheme of Volker and Mehrdad. The proposed scheme corrects the drawback in Volker and Mehrdad’s scheme which is the need of a large amount of storage for storing the secret keys. Security and performance analysis of our scheme proves the proposed scheme to be more efficient and secure.

In this paper we present a platform-level simulation architecture for tactical armored force combat entity agents system by setting up mappings from combat entities to respective combat entity agents. In order to solve the problems on negotiation to enhance overall system efficiency, based on qualitative description of its framework, we place particular emphasis on quantitative analysis. Through transforming the system into a series-wound queueing system, we attain a Markov chain of stationary transition probabilities, since its stationary transition process in negotiation is a discrete state Markov process and accords with real military combat behaviors. Solving the stationary transition equations makes us find high-efficiency negotiation according to optimized system configuration. The obtained results show the effectiveness of the proposed approach and model.

Building intelligent environment is one of crucial challenges for ubiquitous computing developers. To make the environment adapt rationally according to the desire of users, the system should be able to guess users’ interest, by learning users’ behavior, habit or preference. While learning the user preference, dealing with uncertainty and conflict resolution is of the utmost importance. When many users are involved in a ubiquitous environment, the decisions of one user can be affected by the desires of others. This makes learning and prediction of user preference difficult. To address the issue, we propose an approach of user preference learning which can be used widely in context-aware systems. We use Bayesian RN-Metanetwork, a multilevel Bayesian network to model user preference and priority.

Recently semantic web is an important issue in many areas. Ontology specifies the knowledge in a specific domain and defines the concepts of knowledge and the relationships between concepts. To specify and define the knowledge in a specific domain, it is required to generate the ontology which conceptualizes the knowledge. Accordingly, to search the information of potential customers for home-delivery marketing of post office, we design a specific domain to generate the ontology in this paper. And we propose how to retrieve the information, using the generated ontology. We also implement the data search robot which collects the information based on the generated ontology. Therefore we confirm that the ontology and the search robot perform the information retrieval exactly.

In this paper we present an approach where a hybrid computational model is represented as a set of communicating agents composing a multi-agent system. A general concept of representation of connected groups of agents is introduced and utilized for automatic building of schemes to solve a given computational task. We propose a combination of an evolutionary algorithm and a formal logic resolution system which is able to generate and verify new schemes. Furthermore, the adaptive cooperation support of individual computational agents is described, which improves their efficiency in time. These features are implemented within a software system and demonstrated on several examples.

In general, Knowledge base is constructed by domain expert, but in practice, domain expert have no time and energy to do these multifarious tasks. The main objective of a truly user-friendly knowledge base constructing platform is to focus on the decrease of the user’s work. This paper describes the design of a knowledge base constructing platform using multi-agent and XML technology, with which we can easily build a knowledge base, get a veracity-enhanced and structured knowledge base. We firstly describe the main framework, and then illuminate the hierarchical knowledge atoms stored by structured XML Document. All agents, will be described separately in the relevant sections, including mainly control agent, inducting agent, and verifying agent.

This paper describes a unique, open source and free solution to managing workflows in data intensive environments. The Semantic Web offers languages to create ontologies that can provide better semantic matches and help automate workflows. We create an ontology for essay reviewers in various disciplines and use an automated service to query and find comprehensive matches. In particular, we use a free essay review service as a proof-of-concept for our work.

Considering vast and miscellaneous contents in P2P system, intelligent P2P network topology is required to route queries to a relevant subset of peers. Based on the incremental clustering capability of Fuzzy Adaptive Resonance Theory (Fuzzy ART), this paper made use of the modified fuzzy ART to provide small-world P2P construction mechanism, which was not only to categorize peers so that all the peers in a cluster were semantically similar, but, more important, to construct the P2P topology into small-world network structure. In detail, the modified fuzzy ART net was used to cluster peer into one or more appropriate categories according to its data interest, and the reverse selection mechanism in modified fuzzy ART was provided to construct semantic long-range edges among clusters. Simulations demonstrated that P2P small-world network emerged, i. e., highly clustered networks with small diameter, and the information retrieval performance was significantly higher than random topology.

To resolve the problem of agent tending to be helpless when performing tasks in the Internet in which is full of uncertain factors, this paper proposes a hierarchy-like agent assistant service framework that provides mediate services in agent society. This assistance services system is divided into three layers: the bottom is a description layer, which describes agent services and domain ontology by self-defined Ontology Based Knowledge Representation Language (OKRL); the middle is a support layer, which provides management mechanism to organize middle agent nodes and information repository of assistant services for upper layer; the top is assistant services layer, which provides services for registration, advisement and matchmaking of agent services. The growth mechanism and joint matchmaking process of middle agents, which realizes self-organising characteristic for agent assistant system, are also discussed in this paper. Finally, this paper gives evaluation and conclusion.

In this paper, we propose an approach for better ontology interoperability using approximation technology of semantic terminology across multiple ontolgies. We use description logic language for describing ontological information and perform approximate query across multiple ontologies. Meanwhile, we discuss system implementation, and provided some experimental results for evaluating the method.

Dynamic enterprises process optimization (DEPO) is a multi-parametric and multi-objective system optimization problem. This paper proposes a simulation-based process model learning approach for dynamic enterprise process optimization. Some concepts such as Evolving_region, Evolving_Potential, Degenerate_region and Degenerate_limit are proposed to extend the concept of Tabu area. Tabu area extension and connection is successfully presented for realizing rapidly the domain reduction of a candidate set and speeding up global optimization. A distributed parallel optimization environment has been implemented using intelligent agents to validate the proposed approach.

This paper suggests the optimization method of task assigning based on genetic-simulated annealing algorithm focus on the alternative scheme of bidding invitation parameters obtained in the bidding process in MAS. Through reasonable assessing the bidding documents from bid, bidding result determined is obtained, viz. the task assigning scheme. All of the tasks’ assigning optimization is carried out by setting the parameters of genetic-simulated annealing algorithm; the algorithm can search steadily the optimal scheme and overcome the flaws of traditional simulated annealing algorithm such as big undulation and poor astringency. It is found via the analysis of experiment data that the algorithm can improve assignment scheme further rationality and feasibility.

Nowadays amount of information on Internet is dramatically increasing. The ability of facilitating users to achieve useful information is more and more important for search field. CDSE, a model for the domain-based intelligent search engine is proposed in this paper. The model can help users to retrieve what they need by combining text classification with keywords extraction. Several algorithms that use key technologies are proposed, such as statistics, data mining and agents. Then a new criterion named ranking error is contributed to solve the problem of evaluation ranking inefficiency in traditional performance evaluation methodologies. The experimental results indicate that the proposed model can effectively improve retrieval precision and solve the problem of relevant document ranking behind in current search engine.

In this paper, we propose a novel method to create a personalized 3D face in PDA devices from two orthogonal images of users. The approach is useful for making a personalized 3D face keeping identical features of the face and efficient. The basic idea is to transform a canonical Korean male face to individualized ones standing for genuine features of the face by linear transforming filters in the PDA device. Moreover 6 universal facial expressions are simply created just by transformation of the contract values of 18 muscles in a PDA mobile device. Each expression is analyzed with respect to displacement of the feature points from those of the neutral face. Experimental results convey the proposed scheme is quite reliable and efficient with respect to similarity between pictures and the generated 3D face and running time.

In this paper, a skew free Korean character recognition system is developed for PDA devices. There is no doubt text information existing on our real life conveys meaningful messages. It is obviously necessary to capture text images in any place and time in pervasive computing in order to recognize and keep text information in digital forms. In this background, a new mobile-based Korean character recognition system is designed and implemented which is capable of capturing text images and recognizing Korean characters under PDA devices. The algorithm begins with taking images from PDA client followed by skew correction and normalization of character blocks for matching. Experimental results show the proposed scheme is quite novel and efficient.

Our paper proposes an agent-based automated negotiation model. The agents can perform an integrative negotiation with multi-issue in a one-to-many way. The negotiation protocol follows the offer-counteroffer principal, and an adapted mechanism of offer generation strategy. With the utility theory, agent could evaluate the offers and determine the following actions. In order to yield a top-quality deal and shorten the negotiation period, agents propose multiple offers, which consist of a particular combination of issue values and have the identical utility with the given utility. The experiment shows that the model ensures the participants could reach a mutually beneficial agreement in a short time.

In this paper, a multimedia-based decision making model is proposed integrating extended attributed relational graph (eARG) and structured modeling (SM). The proposed methodology manages multimedia information with the use of an extended attributed relational graph. SM technique is adopted to graphically integrate the multimedia information with decision models and hence to support ‘pervasive and seamless decision on demand’: making decision at the time of gazing multimedia objects as if all data and models for decision support are closely at hand.

The petroleum supply chain is an old problem with new challenges. Multi-agent system has recognized as an effective methodology for supply chain management. This paper conducts a systematic multi-agent methodology to China petroleum supply chain system, where multi-agent models are constructed, and the multi-agent system is simulated in Zeus agent platform. Particularly, the advantages of multi-agent modeling and simulation in supply chain management are addressed and demonstrated.

DSAFO (Dynamic Scheduling Agents with Federation Organization) is a novel multi-agent constraint satisfaction algorithm for AGSS problem (a NP-hard scheduling problem). This paper improves on DSAFO by employing a resource requisition strategy, and models this parallel multi-agent algorithm in polyadic π-calculus. The time complexity of the improved DSAFO is O(n3)+O(n2)×t_trans. Experiments show improved DSAFO performs well in AGSS consumptions optimization of resources and man-days. Though it is unstable, improved DSAFO makes good probability to find better solutions than classical heuristics and its distributed and parallel agents viewpoint is potential to deal with distributed dynamic troubles in real applications.

The state of the art statistical machine translation (SMT) systems are based on phrase (a group of words), which are modeled using log-linear maximum entropy framework. In this paper, we constructed a phrase-based statistical machine translation system with additional feature models. The translation model is combined with four specific additional feature functions. When comparing our system with the baseline system of IWSLT2005, we can conclude that our system improve the SMT system accuracy with the same corpus.

We develop an e-learning web application that integrates the materials recommender system to facilitate the learners during the learning process. The system evaluates each learner via the quiz generator by randomly selecting a set of questions that are created by the instructor. Our smart e-learning system helps instructors to create and maintain both compulsory materials and questions. We implemented the system at the faculty of Resource and Environment, Kasetsart University at Sri-racha campus and found that our system got a very good response from the instructors and learners. Furthermore, we propose the global e-learning framework using web service that has an ability to aggregate the recommended materials from other e-learning web sites and predicts more suitable materials to learners.

The learning behaviours of buyers and sellers in the sealed-bid Bargaining Mechanism were studied under the assumption of bounded rationality. The learning process of the agents is modelled by particle swarm optimization (PSO) algorithm. In the proposed model, there are two populations of buyers and sellers with limited computation ability and they were randomly matched to deal repeatedly. The agent’s bidding strategy is assumed to be a linear function of his value of trading item and each agent adjusts his strategy in repeated deals by imitating the most successful member in his population and by own past experience. Such learning pattern by PSO is closer to the behaviours of human beings in real life. Finally, the simulated results show that the bidding strategies of the agents in both populations will converge near the theoretical linear equilibrium solutions (LES).

In this paper, a new efficient hardware architecture and its implementation for an AntNet-based routing are proposed. This architecture is based on the modified AntNet algorithm, and optimized to the ant packets defined in this paper. The modified AntNet is evaluated by the performance comparisons with the original AntNet algorithm through C-level simulations, and then implemented into RTL design. Consequently, the hardware implementation result of the proposed architecture is described.

The Domain Name System (DNS) is the most crucial infrastructure for mapping human-readable host names to the corresponding IP addresses and providing the routing information of Email. Comparing with the top-level domains (TLD) such as the root servers, the local authoritative servers are more vulnerable to device failures and malicious attacks. This paper described the existence condition of authoritative servers and presented a novel domain measurement tool named DNSAuth to collect the information of local authoritative servers automatically. Experiments to the real-life authoritative servers were conducted which highlighting three important aspects: the distribution, the geographic location and their impacts on performance and security. According to five representative attributes, the authoritative servers of China Top100 websites are evaluated and the result shows that only 32% of all the servers act preferably in performance and security.

As a traditional technique of information security, distributed firewall has taken very important position, while problems remain. Correct configuration of distributed firewall policies and keeping individual firewall filter decisions compatible to each other are quite inconvenient for administrators. To realize the comparison between firewalls’ policies, this paper provide FPT(firewall policy tree) model, and the construction algorithm which can turn a firewall policy into a policy tree, as well as the comparison algorithm. Combination of the two algorithms can be used to perform a comparison between distributed firewalls’ policies. By doing this, the paper can obtain the set of data packages on which different firewalls have made inconsistent filter decision, and find out the inconsistency in distributed firewall policies. Besides, this model could be extended to package classification systems for policies comparison.

In this literature, fault detection of an induction motor is carried out using the information of stator current. After preprocessing actual data, Fourier and Wavelet transforms are applied to detect characteristics under the healthy and various faulted conditions. The most reliable phase current among 3-phase currents is selected by the fuzzy entropy. Data are trained with a neural network system, and the fault detection algorithm is carried out under the unknown data. The results of the proposed approach based on Fourier and Wavelet transformations show that the faults are properly classified into six categories.

Failure can be prevented in time by prediction of equipment maintenance so as to promote reliability only if failures can be early predicted. Substantially, it can be boiled down to a pattern recognition problem. Recenty, support vector machine (SVM) becomes a hot technique in this area. When using SVM, how to simultaneously obtain the optimal feature subset and SVM parameters is a crucial problem. This study proposes a method for improving SVM performance in two aspects at one time: feature subset selection and parameter optimization. Fuzzy adaptive particle swarm optimization (FAPSO) is used to optimize both a feature subset and parameters of SVM simultaneously for predictive maintenance. Case analysis shows that this algorithm is scientific and efficient, and adapts to predictive maintenance management for any complicated equipment.

Based on covariance constraint control theory and statistic, the fault predicts problem inner the dynamic complicated system is researched. First of all, according to the important factor to influence the running state by the Multifactor analyze method, we have defined and analyzed the parameter structure of the average and error for state observe values correspondence to the fault point; Simultaneous, the fitting quantitative model and error control threshold of complicated system were presented; secondly, we has given the running principium of integration system to fault detect and designed the network topological structure and the data transfer model; Finally, a simulated example have shown that the technique is effectively and exactly. The theoretical analyze indicated that the presented integration technology of Fault detect has a broad prospect for practical application.

There exists a challenge, i.e., to diagnose failures of such a complex system that has the following characters: (1) it has a causality loop structure; (2) system observed variables are discrete, or continuous, or mixed; and (3) system has time lag, i.e., it has delay of fault effect. For the task, this paper proposes a fault diagnosis prototype system based on causality diagram, and describes the key technologies used. The proposed prototype system comprises datum collection subsystem, model design subsystem, fault diagnosis subsystem, and diagnosis result display subsystem. In this system, the fault knowledge related fault diagnosis is represented in a fault influence propagation diagram (FIPD), which results from a causality diagram. The fault knowledge related delay of fault effect is represented in a 2-time-slice causality diagram (2-TSCD), which is a time stochastic model extended from a FIPD. The method of find fault modes in a FIPD or a 2-TSCD is presented. The proposed prototype system has flexible fault knowledge representation, rapid diagnosis process, and well practicability.

A new FTC scheme based on adaptive radial basis function (RBF) neural network (NN) model for unknown multi-variable dynamic systems is proposed. The scheme designs an adaptive RBF model to built process model and uses extended Kalman filter (EKF) technique to online learn the fault dynamics. Then, a model inversion controller is designed to produce the fault tolerant control (FTC) actions. The proposed scheme is applied to a three-tank process to evaluate the performance of the scheme. The simulation results show that component fault can be quickly compensated so that the system performances are recovered well.

A novel approach for pipeline leak fault detection has been studied, which applies self-organizing fuzzy clustering neural network to identify work status. The proposed method utilized fuzzy neural clustering of DIC method instead of constructing exact mathematical model. After normalizing the sample data, together with prior knowledge, a fuzzy neural network is used to evaluate work status. An adaptive algorithm is developed to diagnose the leak fault. The experiment results have shown the validity and practicability of the method.

High Impedance Faults (HIF) are faults of difficult detection and location while using traditional digital relaying. In this article it is presented a new proposal for detection and location of HIF’s in distribution systems with distributed generation (DG), based on artificial neural networks. The methodology inputs are the local measured voltage and current phase components, supplying as output the detection, classification and location of the fault, when it occurs. The basic characteristics, the algorithm and comparative tests with other detection and location methodologies are presented in this article. The proposed scheme was tested in a simulation platform of a distribution system with DG. The comparative results of the technique with usual fault detection and location schemes show the high efficiency and robustness of the method.

According to the theory of grey relation degree and the definition of the distance of two points, a decision method which is named grey relation close degree model is proposed. The model is successfully used in fault diagnosis of oil-sample analysis. The practical application results show the effectiveness of the proposed approach. It is proved that the method is not only concise but reliable and can greatly widen range of application of grey model.

The increasing need for efficient image storage in hospitals imposes heavy requirements on the design of picture archiving and communication systems. Thus new methods are needed to improve the medical image compression performance. In this paper, we propose an efficient coding algorithm called OSS (Optimal Subband Shift) scheme based on the RB-IWT (Reversible Biorthogonal Integer Wavelet Transform). In the new scheme, the original image is first decomposed by the RB-IWT. Then, the image coefficients of every subband are multiplied by the powers of two. Finally, the SPECK coding is applied. Experimental results show that the OSS scheme does not only provide the PSNR performance better than SPECK using the original RB-IWT without the subband shift, but also has the low coding complexity. So this idea is valuable for future research in medical image coding and its applications.

Fault detection is an important problem in process engineering. A new fault detection method based on artificial immune system is developed for complicated process. Real-valued negative selection algorithm with variable–radius detectors is adopted to generate the detectors set which covers the non-self space. In order to decrease the complexity of detector generation, principal component analysis is introduced to reduce the dimension of the process data. The effectiveness of the proposed method is illustrated by the simulation on the Tennessee Eastman process.

A fault diagnosis system based on integration of rough set theory (RST) and support vector machine (SVM) is developed for induction machine rotor faults detection. The proposed algorithm uses the stator current spectrum as inputs. By RST attribute reduction, redundant attributes are identified and removed. Then the reduction results are used as the input of SVM based classifiers to distinguish different motor conditions. A series of experiments using a three phase 1.5KW induction machine performed in different conditions are used to provide training and test data. The diagnosis results demonstrated that the solution can reduce the cost and raise the efficiency of the diagnosis.

An intelligent process monitoring approach, which consists of process trend recognition and fault detection, is presented in this paper. This method incorporates wavelet transform, symbolic representation of data trend, pattern recognition, and Hidden Markov model (HMM) for intelligent reasoning. A simulation example and an industrial case study have shown the value of this approach.

An approach was put forward to diagnose the multiple faults of rotary machines according to the characteristic frequency spectrum of vibration. Based on the matter-element of extension theory, a matter-element model was built to describe the fault situation of rotary machines qualitatively. The dependent function and degree of a symptom of the fault were introduced to evaluate the possibility of the fault quantitatively. The diagnosis example was taken to validate the approach. The diagnosed result is consistent with the real result.

This paper presents a study on active detection of internal damage based on numerical analysis of vibration dynamics in composites. Damage-induced variations of dynamic parameters are investigated both numerically and experimentally. Finite element method (FEM) is used to compute the modal parameters of composites with or without damages. Natural frequency, modal displacement, modal strain and strain energy are analyzed for the determination of damage severity and location. Vibration measurements are carried out using piezoelectric patch actuators and sensors for comparison and verification of the FEM model proposed in this study. Energy spectrum for wavelet packets decomposition of structural dynamic responses is used to highlight the features of damaged samples. The mechanism of mode-dependent energy dissipation of composite plates due to delamination is revealed for the first time. Both numerical and experimental findings in this study are significant to the establishment of guideline for damage identification in composite structures.

This paper presents a new SoC test scheduling method based on an ant algorithm. The proposed scheduling algorithm formulates the SoC test scheduling problem as a rectangle bin packing problem and uses ACO to cover more solution space to increase the probability of finding optimal solutions. The experimental results conducted using ITC ’02 SoC benchmarks show that the proposed idea gives the test application time comparable to earlier researches in less calculation time.

Fault diagnosis and accommodation(FDA) for nonlinear multi-variables system under multi-fault are investigated in the paper. A complete FDA architecture is proposed by incorporating the intelligent fault tolerant control strategy with a cost-effective fault detection and diagnosis (FDD) scheme based on a multiple-model. The schem efficiently handles the accommodation of both the anticipated and unanticipated failures in online situations. The three-tank with multiple sensor fault concurrence is simulated, the simulating result shows that the fault detection and tolerant control strategy has stronger robustness and tolerant fault ability.

Speaker independent large vocabulary continuous speech recognition technique has always been the research focus in the domain of artificial intelligence and pattern recognition. A Mongolian large vocabulary continuous speech recognition system is introduced in this paper. Mongolian is belonged to Altai phylum, and similar with the Western language. According to the characteristics of Mongolian pronunciation, we build the Mongolian acoustic model. We collected a large size corpus to construct the language model. HTK ( HMM Toolkit ) has been used to construct the system. The experimental results indicated that the design of models related to the Mongolian speech recognition is rational and correct.

Part-of-speech (POS) tagging and shallow parsing are sequence modeling problems. While HMM and other generative models are not the most appropriate for the task of labeling sequential data. Compared with HMM, Maximum Entropy Markov models (MEMM) and other discriminative finite-state models can easily fused more features, however they suffer from the label bias problem. This paper presents a method of Chinese POS tagging and shallow parsing based on conditional random fields (CRF), as new discriminative sequential models, which may incorporate many rich features and well avoid the label bias problem. Moreover, we propose the information feedback from syntactical analysis to lexical analysis, since natural language should be a multi-knowledge interaction in nature. Experiments show that CRF approach achieves 0.70% F-score improvement in POS tagging and 0.67% improvement in shallow parsing. And we also confirm the effectiveness of information feedback to some complicated multi-class words.

The use of an ontology in order to provide a mechanism to enable machine reasoning has continuously increased during the last few years. This paper proposed an automated method for document classification using an ontology, which expresses terminology information and vocabulary contained in Web documents by way of a hierarchical structure. Ontology-based document classification involves determining document features that represent the Web documents most accurately, and classifying them into the most appropriate categories after analyzing their contents by using at least two pre-defined categories per given document features. In this paper, Web documents are classified in real time not with experimental data or a learning process, but by similarity calculations between the terminology information extracted from Web documents and ontology categories. This results in a more accurate document classification since the meanings and relationships unique to each document are determined.

In recent years, several methods on human emotion recognition have been published. In this paper, we proposed a scheme that applied the emotion classification technique for emotion recognition. The emotion classification model is Support Vector Machines (SVMs). The SVMs have become an increasingly popular tool for machine learning tasks involving classification, regression or novelty detection. The Emotion Recognition System will be recognise emotion from the sentence that was inputted from the keyboard. The training set and testing set were constructed to verify the effect of this model. Experiments showed that this method could achieve better results in practice. The result showed that this method has potential in the emotion recognition field.

In recent years, several methods on human emotion recognition have been published. In this paper, we proposed a scheme that applied the emotion classification technique for emotion recognition. The emotion classification model is Support Vector Machines (SVMs). The SVMs have become an increasingly popular tool for machine learning tasks involving classification, regression or novelty detection. The Emotion Recognition System will be recognise emotion from the sentence that was inputted from the keyboard. The training set and testing set were constructed to verify the effect of this model. Experiments showed that this method could achieve better results in practice. The result showed that this method has potential in the emotion recognition field.

Environments based on ambient intelligence require new interfaces that allow a natural interaction. The development of these interfaces has to be done in a standard way, considering the dynamic characteristics of these environments. In this paper we present the results in the development of an intelligent environment and a description language for the automatic generation of a spoken dialogue interface, which adapts to the characteristics of every environment.

New applications of artificial neural networks are capable of recognition and verification of effects and the safety of commands given by the operator of the technological device. In this paper, a review of selected issues is carried out in relation to estimation of results and safety of the operator’s commands as well as the supervision of the process. A view is offered of the complexity of effect analysis and safety assessment of commands given by the operator using neural networks. The first part of the paper introduces a new concept of modern supervising systems of the process using a natural language human-machine interface and discusses general topics and issues. The second part is devoted to a discussion of more specific topics of automatic command verification that has led to interesting new approaches and techniques.

Many traditional algorithms use single metric generated by multi-events to detect intrusion by comparison with a certain threshold. In this paper we present a metric vector-based algorithm to detect intrusion while introducing the sample distance for both discrete and continuous data in order to improve the algorithm on heterogeneous dataset. Experiments on MIT lab Data show that the proposed algorithm is effective and efficient.

Service robots are built to help ordinary people in their daily life. The most desirable way for communicating with service robots is through natural language interfaces. In our research, we develop a novel bio-inspired approach, called the collaborative behavior based approach, to build a natural language interface between a robot and its human user. In this approach, knowledge about collaborative behaviors of both the robot and its user is applied to solve ambiguity in a natural language. In building a system that is based on knowledge about behaviors, a key issue is representing the knowledge. So far, little research has been done in representing knowledge about behaviors. In this paper, we describe the collaborative behavior based approach, with emphasis on its knowledge representation structures.

Word segmentation and part of speech (POS) tagging are basis of processing oracle-bone inscription by using computer. It is hard to build a large tagged oracle-bone inscription corpus with grammar information. This is an obstacle if we want to use statistical method. In this paper, we propose to solve both problems with methods combining corpus-based and rule-based approaches. The accuracy of segmentor and tagger are 98.33% and 96.75% respectively. Our experiment result shows that the combining method is quite practical for processing the oracle-bone inscription, especially when the corpus is too sparse. In the end, we briefly discuss how to use the tagged result to complete syntax analysis with rule-based method.

Nowadays technological devices can already be provided with enough intelligence to understand and act appropriately on voice commands. The voice communication with technological devices becomes a stronger challenge as technology becomes more advanced and complex. In this paper, a natural language interface is presented which consists of the intelligent mechanisms of human identification, speech recognition, word and command recognition, command syntax and result analysis, command safety assessment, technological process supervision as well as human reaction assessment. In this paper, a review is carried out of selected issues with regards to recognition of speech commands in natural language given by the operator of the technological device. A view is offered of the complexity of the recognition process of the operator’s words and commands using neural networks made up of a few layers of neurons. The paper presents research results of speech recognition and automatic recognition of commands in natural language using artificial neural networks.

In this paper, we present a new approach to align sentences in bilingual parallel corpora based on the use of the linguistic information of the text pair in Gaussian mixture model (GMM) classifier. A feature parameter vector is extracted from the text pair under consideration. This vector contains text features such as length, punctuation score, cognate score and a bilingual lexicon extracted from the parallel corpus under consideration. A set of manually prepared training data has been assigned to train the Gaussian mixture model. Another set of data was used for testing. Using the Gaussian mixture model approach, we could achieve error reduction of 160% over length based approach when applied on English-Arabic parallel documents. In addition, the results of (GMM) outperform the results of the combined model which exploits length, punctuation, cognate and bilingual lexicon in a dynamic framework.

We demonstrate that European Asian options with harmonic averaging behave better than Asian options with arithmetic and geometric averaging procedures under some situation. Approximation methods for the valuation of harmonic average options and numerical illustrations are also given.

The problem of research cost estimation is a typical multi-factors estimation issue, which has not been solved satisfactorily. A method integrating rough sets theory and artificial neural network is presented to estimate cost. In term of the important degree of input influencing factor to output, rough set approach and the conception of information entropy are employed to reduce the parameters of the input parameter set with no changing classification quality of samples. Thus, the number of the input variables and neurons is gotten, and the cost estimation model based on rough set and BP artificial network is set by learning from the original data of typical samples. At last, its application to the cost estimation of missile system is given. It was shown that the approach can reduce the training time, improve the learning efficiency, enhance the predication accuracy, and be feasible and effective.

An adaptive neuro-fuzzy based method for analyzing odour generation factors to the perception of livestock farm odour was proposed. In this approach, the parameters associated with a given membership function could be tuned so as to tailor the membership functions to the input/output data in order to account for these types of variations in the data values. A multi-factor livestock farm odour model was developed, and both numeric factors and linguistic factors were considered. The proposed method was tested with a livestock farm odour database. The results demonstrated the effectiveness of the proposed approach in comparison to a typical neural network.

According to the analysis of credit risk assessment in commercial banks, a set of index system is established. The index system combines financial factors with non-financial factors for credit risk assessment. The credit rating is separated into five classes- normality, attention, sublevel, doubt and loss. To classify the credit risks of five classes, a multi-layer support vector machines (SVM) classifier is established to assess the credit risk. In order to verify the effectiveness of the method, a real case is given and BP neural network is also used to assess the same data. The experiment results show that multi-layer SVM classifier is effective in credit risk assessment and achieves better performance than BP neural network.

By characterizing the number of events occurred in any time as a Poisson distributed random fuzzy variable, and dealing with the intensity function of Poisson process with time-dependent intensity as a fuzzy variable, the concept of a Poisson process with fuzzy time-dependent intensity is defined. And then, related to the process, several theorems on the expected value of the number of events occurred in any time and average chances of random fuzzy events are respectively discussed.

Random fuzzy theory is introduced to evaluate the option value of R&D project in this paper. Based on a new theoretical insight, we develop a random fuzzy jump amplitude model to price R&D option. In classic stochastic jump amplitude model, the value of R&D option depends on the expected number of jumps and the expected size of the jumps. However, in practice, it is so hard to get enough data that the distribution functions of them are difficult to be confirmed accurately. So random fuzzy theory can deal with these problems better. In this paper, the number of jumps and the size of jumps are taken as random fuzzy variables, and investment costs and future cash flows are depicted as fuzzy variables. New research tools for option pricing of R&D project are developed and the capability of dealing with practical problems is enhanced.

The purpose of this paper is to introduce a two-person zero-sum matrix game in which the payoffs are characterized as fuzzy random variables. Based on fuzzy random expected value operator, a fuzzy random expected minimax equilibrium strategy to the game is defined and the existence of the strategy is proved. Then an iterative algorithm based on fuzzy random simulation is designed to seek the minimax equilibrium strategy. Finally, a numerical example is provided to illustrate the effectiveness of the algorithm.

This paper develops a fuzzy model that considers the uncertain attributes of change orders in a research and development (R&D) project. It is assumed that the arrival times of change orders follow a fuzzy renewal process. The contingency based on the confidence level offered by the decision-maker is estimated by fuzzy simulation. Furthermore, the effect of schedule delays which increase further the project cost and schedule is considered explicitly. Finally, three numerical examples are presented to illustrate the application of the proposed model.

Many actual project problems generally belong to large-scale TSP, The large-scale TSP as a famous NP-hard problem will be faced with the dual challenges of the optimization performance and the CPU run-time performance for any single algorithm. In fact, the optimal solution is not pursued overwhelmingly in actual projects, but it needs to meet certain optimization efficiency. This paper reduces the problem’s complexity based on the idea of “divide and rule”. We use the method of K-Means cluster to divide the nearest neighbor quickly. Then, we employ parallel computing method to all divided areas by using combination of genetic algorithm and ant algorithm (GAAA). Finally, we globally link all the subsets using the method of K centers connect. The results of simulations show that its complexity has been greatly reduced and we can quickly obtain a satisfactory solution to the large-scale problem. It is one effective way to solve the large-scale complex problems.

Fourteen properties and their interdependencies contributed to the analysis of six classes implication operators–S-implications, R-implications, QL-implications, Force implication, f- generated implication operator and g- generated implication operator– are explored in this paper. It is found that all the properties can be inferred from three mutually independent properties. Then, the proven concerning which properties are true, false or satisfied for each of the six classes of implication operators is given. Based on which, the author get the property I(x(n(x)) = n (x) for all x ∈[0,1] that does not hold for force implication, f-generated implication operator and g-generated implication operator.

Any quantum algorithm has to be implemented by quantum circuit. However, due to the existence of decoherence time and the difficulty of importing ancillae qubits, we introduce a new method which can greatly reduce the operating time steps by making good use of relaxing qubits without new ancillae qubits.This concept presented in this paper can be generalized to the other quantum networks.

This paper mainly focuses on the modeling of oil well pressure data compensation system(OWPCS) based on Neural Networks(NN). Firstly, the operational principle and configuration of OWPCS is described. Then the currently widely used modeling method for OWPCS is given, and its limitations and disadvantages are also illustrated. Secondly, in order to solve the OWPCS modeling problem more reasonably, a new approach based on Neural Network is proposed. Thirdly, the feasibility of using NN to solve this problem is analyzed, and a three-layer BP network is constructed to testify the new modeling method. Fourthly, considering the defect of BP learning algorithm and the special application environment of OWPCS, some improvements are given. Finally, experiment results are presented to show the reasonableness and effectiveness of the new method.

In this paper, we present an approach for moving object contour tracking in video by using an improved dual-front active contour model. Dual-front active contour model is first proposed for medical image segmentation. In order to adapt it to object tracking problem, we make two improvements on the original model. First, region force of the external front is modified by restricting its support region. This modification can speed up the algorithm greatly but may result in the active contour’s wrong convergence to the real object boundary when it locates in a large homogeneous region. Then, a new function called quasi-balloon force is brought into the model by modifying its active region construction method. It can not only solve the problem result from the first improvement but also make tracking more flexible. The algorithm does not need an a priori shape so it is fit for deformable object tracking. By adjusting the parameters, it can be used to track fast moving target. Since the level set method is used, the topology change of the object can be controlled automatically. And no static background of the scene is assumed which means the contour can be tracked under the condition that both the camera and the object are moving. Experimental results demonstrate its effectiveness and robustness.

Inspection robot must detect the obstacles from the complex background according to their types when it is crawling along the power transmission line in order to negotiate reliably. In ideal cases, robot’s vision system can give satisfactory results, however, motion blur due to camera motion caused by wind or other unknown causes can significantly degrade the quality of the image acquired. It is an undesired effect. In this paper, a complete motion deblurring procedure for obstacle images has been proposed, we try to analyze the running environment of the robot to develop the model of the motion blur. The acquired motion blur information is used to identify the point spread function (PSF) as well as restore the blurred image at the same time. Experiments on real blurred images on power transmission line prove the feasibility and reliability of this algorithm.

In this paper, we suggest a viewpoint-invariant face recognition model based on view-based representation. The suggested model has four stages: view-based representation, viewpoint classification, frontal face estimation and face recognition. For view-based representation, we obtained the feature space by using independent subspace analysis, the bases of which are grouped like the neurons in the brain’s visual area. The viewpoint of a facial image can be easily classified by a single-layer perceptron due to view-dependent activation characteristic of the feature space. To estimate the independent subspace analysis representation of frontal face, a radial basis neural network learns to generalize the relation of the bases between two viewpoints. Face recognition relies on a normalized correlation for selecting the most similar frontal faces in a gallery. Through our face recognition experiment on XM2VTS [9], we obtained a face recognition rate of 89.33%.

In this paper, we propose a new scheme based on contourlet to represent the visual information from the aspect of magnitude and orientation. Based on the detail statistics analysis of the individual, joint behaviors and correlations of contourlet coefficients of natural images across scales, positions and directions, it reveals strong local dependencies and clustering when the coefficients are at low amplitude. According to these fundamental findings, a novel embedded block with significance selecting model is developed to present the transformed coefficients. Experimental results demonstrate that our proposed representation is efficient. It is comparable to the wavelet coder in terms of the PSNR metric, and visually superior to the wavelet coder for the images with detailed texture, which is more fit for the Human Visual System.

Inspection robot must plan its behavior to negotiate obstacles according to their types when it is crawling along the power transmission line. For this purpose, a visual navigation system is designed to recognize the obstacles and locate their positions by stereovision.We propose a structure-constrained obstacle recognition algorithm based on improved circle detection methods to recognize obstacles from complex background robustly. After the obstacle is recognized, a region based stereo matching algorithm is used to search the correspondence points in the stereo images, and the position of the obstacle relative to the robot is calculated by 3D reconstruction. Experiments with simulation and real transmission line show its effectiveness.

In this paper we present a semantic analyzer for aiding emotion recognition in Chinese. The analyzer uses a decision tree to assign semantic dependency relations between headwords and modifiers. It is able to achieve an accuracy of 83.5%. The semantic information is combined with rules for Chinese verbs containing emotion to describe the emotion of the people in the sentence. The rules give information on how to assign emotion to agents, receivers, etc. depending on the verb in the sentence.

The approach of emotion estimation from the conventional text was for estimating superficial emotion expression mainly. However emotions may be included in human’s utterance even if emotion expressions are not in it. In this paper, we proposed an emotion estimation algorithm for conversation sentence. We gave the rules of emotion occurrence to 1616 sentence patterns. In addition, we developed a dictionary which consisted of emotional words and emotional idioms. The proposed method can estimate emotions in a sentence by matching the sentence pattern of emotion occurrence and the rule. Furthermore, we can get two or more emotions included in the sentence by calculating emotion parameter. We constructed the experiment system based on the proposed method for evaluation. We analyzed weblog data including 253 sentences by the system, and conducted the experiment to evaluate emotion estimation accuracy. As a result, we obtained the estimation accuracy of about 60 %.

The paper presents our recent work on the acoustic and physiological feature analysis of affective speech. An affective speech corpus is first built up. It contains passages read in neutral state and ten typical emotional states selected in Pleasure Arousal Dominance (PAD) space. Physiological data, including electrocardiogram, respiration, electro dermal data, and finger pulse, are also collected synchronized with speech. Then, based on the corpus, the relationship between emotional categories\dimensions and acoustic\physiological features is analyzed in three methods: average, correlation and co-clustering. The analysis results show that most acoustic features and physiological features are significantly correlated with the arousal dimension, whereas respiration features are more correlated with the pleasure dimension.

Authors of news stories through their choice in words and phrasing inject an underlying emotion into their stories. A story about the same event or person can have radically different emotions depending on the author, newspaper, and nationality. In this paper we propose a system to judge the emotion of a news article based on emotion word, idiom and modifier dictionaries. This type of system allows one to judge the world opinion on varying topics by looking at the emotion used within news articles about the topic.

In this paper, we build a Japanese emotion corpus and perform statistical analysis on it. We manually entered in about 1,200 example dialogue sentences. We collected statistical information from the corpus to analyze the way emotion is expressed in Japanese dialogue. Such statistics should prove useful for dealing with emotion in natural language. We believe the collected statistics accurately describe emotion in Japanese dialogue.

Quantifiers and numerals often cause mistakes in Chinese-Japanese machine translation. In this paper, an approach is proposed based on the syntactic features after classification. Using the difference in type and position of quantifiers between Chinese and Japanese, quantifier translation rules were acquired. Evaluation was conducted using the acquired translation rules. Finally, the adaptability of the experimental data was verified and the methods achieved the accuracy of 90.75%, which showed that they were effective in processing quantifiers and numerals.

Since the completeness of the concept lattice, its construction efficiency is a key of restricting the application. For the inevitable redundant information occurred in the construction process of partitioned concept lattice, an improved incremental construction algorithm PHCL of horizontal partitioned concept lattice, called the pruning based incremental algorithm is proposed, which uses a pruning process to eliminate redundant information during the construction, to improve the construction efficiency of horizontal partitioned concept lattice. In the end, the experiment results prove the correctness and validity of the pruning based algorithm PHCL by taking the star spectra from the LAMOST project as the formal context.

This paper proposes spatial outliers detection method of studying multiple non-spatial attributes based on special objects. The spatial outliers detection algorithm based on the Mahalanobis distance is proposed in this paper. The simulated experiment results demonstrate this method is feasible and effective, simultaneously the time complexity of center algorithm is analysed. Except the research type mentioned, spatial outliers detection still includes time-order data and time-space data outliers detections, which are related to the attribute values of other neighbors.

The similarity measure is derived using fuzzy entropy and distance measure. By the relations of fuzzy entropy, distance measure, and similarity measure, we first obtain the fuzzy entropy. And with both fuzzy entropy and distance measure, similarity measure is obtained. We verify that the proposed measure become the similarity measure.

Software metrics are collected at various phases of the software development process. These metrics contain the information of software and can be used to predict software quality in the early stage of software life cycle. Intelligent computing techniques such as data mining can be applied in the study of software quality by analyzing software metrics. Clustering analysis, which is one of data mining techniques, is adopted to build the software quality prediction models in early period of software testing. In this paper, three clustering methods, k-means, fuzzy c-means and Gaussian mixture model, are investigated for the analysis of two real-world software metric datasets. The experiment results show that the best method in predicting software quality is dependent on practical dataset, and clustering analysis technique has advantages in software quality prediction since it can be used in the case having little prior knowledge.

Vehicle routing problems (VRP) arise in many real-life applications within transportation and logistics. This paper considers vehicle routing models with fuzzy travel times and its hybrid intelligent algorithm. Two new types of credibility programming models including fuzzy chance-constrained programming and fuzzy chance-constrained goal programming are presented to model fuzzy VRP. A hybrid intelligent algorithm based on fuzzy simulation and genetic algorithm is designed to solve the proposed fuzzy VRP models. Moreover, some numerical experiments are provided to demonstrate the applications of the models and the computational efficiency of the proposed approach.

An ant colony optimization algorithm is designed to solve continuous optimization models. Based on this algorithm, a hybrid intelligent algorithm combined with fuzzy simulation and neural network is introduced for solving fuzzy chance constrained models. Several numerical examples are given to show the algorithms effective. As an application, a fuzzy inventory model is established and solved with the hybrid intelligent algorithm.

This paper is devoted to the construction of infinite dimensional product possibility space as well as its applications in theory of fuzzy processes. First, the countably infinite dimensional product ample field, and the extension of countably many product possibility measures based on a continuous triangular norm are discussed. Then the results are generalized to the case of uncountably many factors. Finally, the obtained results about the product possibility space is applied to the construction of a fuzzy vector, a sequence of fuzzy variables and a fuzzy process.

This paper presents a new methodology for designing a detection, isolation, and identification scheme for sensor faults in linear time-varying systems. Practically important is that the proposed methodology is constructed on the basis of historical data and does not require a priori information to isolate and identify sensor faults. This is achieved by identifying a state space model and designing a fault isolation and identification filter. To address time-varying process behavior, the state space model and fault reconstruction filter are updated using a two-time-scale approach. Fault identification takes place at a higher frequency than the adaptation of the monitoring scheme. To demonstrate the utility of the new scheme, the paper evaluates its performance using simulations of a LTI system and a chemical process with time-varying parameters and industrial data from a debutanizer and a melter process.

The Open Vehicle Routing Problem was brought forward several decades ago, but it has still received little attention from researchers for a satisfactory solution. In this paper, a novel real number encoding method of Particle Swarm Optimization (PSO) for Open Vehicle Routing Problem is proposed. The vehicle is mapped into the integer part of the real number; and the sequence of customers in the vehicle is mapped into the decimal fraction of the real number. After decoding, several heurist methods are applied into the post-optimization procedure, such as Nearest Insertion algorithm, GENI algorithm, and 2-Opt. They are used to optimize the inner or outer routes and modify illegal solutions. In the experiments, a number of numerical examples are carried out for testing and verification. The performance of the proposed post-optimization algorithm is analyzed and the particle swarm optimization algorithm is compared with other heuristic methods for the same problem.

Facility layout problems concerning space layout optimization have been investigated in depth by researchers in many engineering fields. In this paper, a particular facility layout problem with aisles and two objectives at minimizing total cost of material handling and maximizing adjacent requirement between resources is discussed and formulated as an nonlinear mixed-integer programming. To solve the NP-hard problem, a multiple objective genetic algorithm approach with local search method is developed. The application on a practical FLP case and numerical analysis show the effectiveness and efficiency of the proposed method on FLPs.

As one of the most important management strategies, supply chain management (SCM) is increasingly being emphasized. And, an increasing focus is placed on the integration of overall supply chain resources. In running the strategy, one of key problems is how to judge the suitability that a project is managed by SCM paradigm. For this reason, a novel feasibility analysis approach based on game theory is presented for running a supply chain project. First of all, some basic conditions for judging the feasibility of a SCM project are discussed both on individual rationality and group rationality. Then, a viable bargain price range of candidate partners is proposed by using Bayes-Nash equilibrium, and a numerical example is given to illustrate its application. Finally, we discuss the relationship between the bargain price range and the competitive index in a SCM project.

Industry clusters with urbanization interactive development is a multi-factor complex system. In this research, we present a system dynamics methodology to construct an industrial clustering and urbanization interactive dynamics model, and further to make its simulation computation and analysis. The simulation results show that the development of industrial clusters stimulates the rise of urbanization level, and the level of urbanization and the growth of industrial clusters promote the development of industrial clusters. Also, a case study for Yiwu City, P.R.China is conducted. It shows that the urbanization of Yiwu City is a fairly advanced virtuous circle.

Numerical examples and some simple systems, for instance an inverted pendulum on a cart, are often used in the analytical and simulation study of Networked Control Systems (NCSs). In this paper, we propose to use a system of “Two inverted pendulums Coupled by a Spring” to extend the existing study to investigate: (1) asynchronised multi-rate sampling in a NCS, (2) random, distributed and discrete time delay in networked signal-transfer with possible loss of signals, (3) NCS for nonlinear systems, (4) Robust control of NCS with plant uncertainties and external disturbances, and (5) using NCS to implement controller structure beyond the limit of decentralised control. This system has been used, in the context of point-to-point connections, to study a number of control problems including a study on robust stabilisation of nonlinear systems via decentralised control. Therefore the performances of the system having a NCS structure can be compared with those having a traditional structure of fixed connections. The detailed model, some simulation results of the system under both of a traditional structure and a NCS structure, and proposed further studies are presented in the paper.

In this paper, a class of adaptive control systems is to be studied, in which the identification loop is closed over communication networks. The network-induced delays caused by the communication network are then inevitable and randomly time-varying in general. And these delays between controller and identifier can make a mess of the transmitted packet sequences which may deteriorate the control performance even destabilize the system. Obviously, the randomly time-varying delay is a key problem in networked adaptive control systems, and a strategy using the concept of fixed maximum delay is presented to avoid this problem. The adaptive system is proven to be convergent if such strategy is to be used. After that, with a view to further improve the performance, another modified strategy is proposed and the simulation results verify its validity and practicability.

In this paper, the statistical model of communication network delay based on the data measured from a network in operation is studied. In our study, it is found that modeling the round trip time (RTT) in our network by a single statistical model is not adequate. Therefore two combined statistical distributions, Pareto distribution and generalized exponential distribution, are used to develop our model. This approach is verified by the standard Chi-square test widely used by statisticians. Based on the model developed, a gain-scheduling algorithm to adjust controller gain to compensate time delay in a networked control system (NCS) is developed. This paper presents the results of our initial study. In addition, further research based on the design methodology pioneered in this paper is also proposed.

This paper provides an overview of the current field in wireless networks for monitoring and control. Alternative wireless technologies are introduced, together with current typical industrial applications. The focus then shifts to wireless Ethernet and the specialised requirements for wireless networked control systems (WNCS) are discussed. This is followed by a brief look at some current WNCS research, including reduced communication control.

A good message scheduling algorithm could give timeliness guarantee to real-time control network. Based on DM algorithm, this paper presents an improved deadline-based algorithm, Deadline Monotonic with Urgent Message Considered (DMUMC), for FF control network message scheduling. The main idea of DMUMC is that deadline of urgent aperiodic message, such as alarms, was taken into account when establishing BAT. The simulation results show that DMUMC algorithm significantly reduces response time of urgent aperiodic message by adjusting service time of low priority periodic message with large deadline, and improves timeliness and message schedulability of FF control network, compared with traditional methods that didn’t consider the issue of aperiodic message.

Communication in an automobile is coming to be important and complex little by little, because of hundreds of circuits, sensors, and many other electrical components. For example, some high-end luxury cars contain more than three miles and nearly 200 pounds of wiring. Therefore, the difficulty in connecting these components is gradually important issue for the automotive industries. As a solution to this affair, networking provides a more efficient method for today’s complex in-vehicle communications. And then, many automotive buses have emerged in last years. Especially, to connect multimedia devices(eg, audio, GPS navigation system, DVD player, PC), the MOST(media oriented systems transport) is proposed to automotive industries as multimedia network in an automobile. To follow the tendency of network integration, we try to connect CAN(controller area network) to MOST. Indeed, it is that we use MOST as control network instead of CAN. To do this, we explain the characteristic of MOST and CAN. Secondly, the inter-operability unit is analyzed by queueing model. Thirdly, we substitute the real CAN traffic for queueing model.

Owing to the fact that the networked control system is featured by uncertain delays, the stochastic time delay can be transformed into a deterministic delay by placing a special mount of buffers at the nodes in the networked control system, thus transform the stochastic networked control system into a deterministic delay system. A controller applying the predictive functional control is hence designed, for the purpose of improving the control capability through realizing the presumption of mould matching and multi-step prognosis. The results of both the simulation based on MATLAB and the experiment based on distributed control system (DCS) show that the method put forward in this paper are not only correct but also effective.

The Fuzzy C-Means (FCM) algorithm is commonly used for clustering. It is one of the problems in association rules mining that a great number of rules generated from the dataset makes it difficult to analyze and use. From the angle of knowledge management, a modified FCM algorithm is proposed and applied to association rules clustering, which partitions these rules into the given classes by the attribute’s weight based on information gain for evaluating the attribute’s importance. Experiment with the UCI dataset shows that this algorithm can efficiently cluster the association rules for a user to understand.

In order to meet the require of high speed and comfort and low producing cost, semi-active suspension for high-speed railway vehicle is adopted to control the vibration of car body. But accurate mathematic model of vehicle dynamics system is difficult to establish for the particularity of wheel-rail interaction and non-linear of system sky damper control needs the absolute speed of car body that its exact value is difficult to obtain. So adaptive fuzzy control method based on the acceleration feedback is put forward in the paper. The method which can modify automatically the scaling factor and control rules according to the acceleration and its change rate of car body is simulated by using ADAMS and MATLAB software. The simulation results show that the method can attenuate the vibration of car body and improve the comfort and stationarity effectively.

Car plate Localization, which remains a difficult problem under complicated environment, is the key problem in many traffic related applications. In this paper we describe a new method based on modified Pulse Coupled Neural Network (PCNN) with adaptive threshold, which can capture relatively complete objects in human perception. After inverse filtering, PCNN processing is applied to produce a firing time sequence image. Then car plates’ position and rotated angle can be extracted from the firing image. Experiment results show that the correct car plate locating rate reaches 98%, which is higher than other Localization methods on the same image database.

Having implemented discrete stationary wavelet transform (DSWT) to an image, combining generalized cross validation (GCV), noise is reduced directly in the high frequency sub-bands which are at the better resolution levels and local contrast is enhanced by combining de-noising method with non-linear gain operator (NGO) in the high frequency sub-bands which are at the worse resolution levels. In order to enhance the global contrast for the image, the low frequency sub-band image is also enhanced employing in-complete Beta transform (IBT) and simulated annealing algorithm (SA). IBT is used to obtain non-linear gray transform curve. Transform parameters are determined by SA so as to obtain optimal non-linear gray transform parameters. In order to avoid the expensive time for traditional contrast enhancement algorithms, which search optimal gray transform parameters in the whole gray transform parameters space, a new criterion is proposed with gray level histogram. Contrast type for original image is determined employing the new criterion. Gray transform parameters space is given respectively according to different contrast types, which shrinks gray transform parameters space greatly. Finally, the quality of enhanced image is evaluated by a total cost criterion. Experimental results show that the new algorithm can improve greatly the global and local contrast for an image while reducing efficiently gauss white noise (GWN) in the image. The new algorithm is more excellent in performance than histogram equalization, un-sharpened mask algorithm, WYQ algorithm and GWP algorithm.

In the design of standalone hybrid wind/photovoltaic power systems, the optimal sizing is an important and challenging task. The coordination among renewable energy resources, generators, energy storages and loads is very complicated. The size of wind turbine generators (WTGs), the size of photovoltaic (PV) panels and the capacity of batteries must be optimized when sizing a standalone hybrid wind/PV power system, which is formulated as a nonlinear integer programming problem. Our objective is selected as minimizing the total capital cost, subject to the constraint of the Loss of Power Supply Probability (LPSP) calculated by simulation. We propose a specific Graph-based Ant System to solve the concerned problem intuitively and easily. The death penalty method is used to deal with the constraint. Simulations have shown that the proposed Graph-based Ant System is efficient with respect to the quality of solutions and computing time.

We propose a new approach to tackle the manpower planning problem with multiple types of jobs in a long planning horizon, where dynamic demands for manpower must be fulfilled by allocating enough number of employees with qualified skills. We first apply Lagrangean relaxation to decompose the problem into a number of subproblems, each corresponding to one skill type, and then develop a coordination scheme based on a Genetic algorithm, which updates the Lagrangean multipliers to maximize the dual objective function. We report computational results, which demonstrate the effectiveness of our approach.

A multistage stochastic financial optimization manages portfolio in constantly changing financial markets by periodically rebalancing the asset portfolio to achieve return maximization and/or risk minimization. In this paper, we present a decision-making process that uses our proposed Quantum-behaved Particle Swarm Optimization (QPSO) Algorithm to solve multi-stage portfolio optimization problem. The objective function is classical return-variance function. The performance of our algorithm is demonstrated by optimizing the allocation of cash and various stocks in S&P 100 index. Experiments are conducted to compare performance of the portfolios optimized by different objective functions with Particle Swarm Optimization (PSO) algorithm and Genetic Algorithm (GA) in terms of efficient frontiers.

Utterance verification is a process, in which a spoken utterance is verified against the given keyword. This process is used to make a decision on acceptance or rejection. In this paper, we propose a new approach to the utterance verification, using a boosting classifier with ten confidence measures. This classifier combines a set of ’weak’ learners into a ’strong’ one. The experimental results present that it can remarkably improve the verification performance. Compared with a single confidence measure, the equal error rate is reduced by up to 23%. The results also show that the boosting classifier is better than the SVM and MLP classifiers, in term of the equal error rate.

Artificial Neural Networks (ANNs) have been studied intensively in the field of computer science in recent years and have been shown to be a powerful tool for a variety of data-classification and pattern-recognition tasks. In this work, computerised diagnostic performance of hepatitis disease was investigated by various ANNs. Multilayer Perceptron, Radial Basis Function Neural Network, Conic Section Function Neural Network, Probabilistic Neural Network, and General Regression Neural Network structures have been used for this purpose. To determine diagnostic performance of networks for hepatitis disease, cross validation method and ROC analysis were applied.

Information extraction is becoming an important task due to the vast growth of the online texts. Pattern rule induction is one kind of main methods to do information extraction. Manually constructing pattern rules is tedious and error prone. In this paper, we present GRID_CoTrain, a weakly supervised paradigm by bootstrapping GRID (a supervised rule induction system) with co-training and active learning. We also utilize external knowledge resource such as WordNet and existing ontology knowledge to optimize the learned pattern rules.

This paper presents a benign prostatic hyperplasia management system which allows patients to cut down the number of hospital visits by using mobile devices like PDA phones or using Web applications and some information management techniques for patient care.

In this paper, a system based on Artificial Neural Network (ANN) for classifying automobile parts with different shapes is presented. The system gets the original information from an image sensor, classifies two sorts of automobile parts with different shapes after processing these images. The classifier designed in this paper adopts the ANN with an improved BP algorithm. The perimeter, acreage and the degree of the decentralization of the automobile parts in the intensity image are taken as the input feature vectors. For the No.1 part and No.2 part, the experimental result indicates that the recognition accuracy rate can reach up to 99% in this system and the productivity will be improved obviously comparing with the checking result in the off-line system manually.

An approach for solving uncorrelated optimal discriminant vectors (UODV), called indirect uncorrelated linear discriminant analysis(IULDA), is proposed. This is done by establishing a relation between canonical correlation analysis (CCA) and Fisher linear discriminant analysis(FLDA). The advantages of our method for solving the UODV over the two existing methods are analyzed theoretically. Experimental result based on the Concordia University CENPARMI handwritten character database has shown that our algorithm can increase the recognition rate and the speed of feature extraction.

Collaborative filtering systems based on a matrix are effective in recommending items to users. However, these systems suffer from the fact that they decrease the accuracy of recommendations, recognized specifically as the sparsity and the first rater problems. This paper proposes the constructing full matrix through Naïve Bayesian, to solve the problems of collaborative filtering. The proposed approach uses Naïve Bayesian, in order to convert the sparse ratings matrix into a full ratings matrix; subsequently using collaborative filtering, to provide recommendations. The proposed method is evaluated in the EachMovie dataset and the approach is demonstrated to perform better than both collaborative filtering and content-based filtering.

A novel particle filter, enhancing particle swarm optimization based particle filter (EPSOPF), is proposed for visual tracking. Particle filter (PF) is sequential Monte Carlo simulation based on particle set representations of probability densities, which can be applied to visual tracking. However, PF has the impoverishment phenomenon which limits its application. To improve the performance of PF, particle swarm optimization with mutation operator is introduced to form new filtering, in which mutation operator maintain multiple modes of particle set and optimization-seeking procedure drives particles to their neighboring maximum of the posterior. When applied to visual tracking, the proposed approach can realize more efficient function than PF.

This paper intends to cope with single target tracking nonlinear filtering problem with an Adaptive Neuro-Fuzzy Inference System (ANFIS). ANFIS provides faster adaptation, adequate convergence and easy using over other standard filters. The ANFIS used in this study is trained on-line while the target is moving to estimate the next position of target at the end of the training. The proposed system calculates the speed and the acceleration rate of the object and estimates the next absolute position of the target between two position measurements interval. Estimation performance of the presented system has been tested using few predetermined position data. The test results show that the proposed ANFIS position estimator system has been successively estimated the next position of the moving target and can be used in real target tracking systems.

In emergent airdropping, there are some special requirements for the vehicle routing problem. The airdropping personnel and equipment, geographically scattered, should be delivered simultaneously to an assigned place by a fleet of vehicles as soon as possible. In this case the objective is not to minimize the total distance traveled or the number of vehicles as usual, but to minimize the time of all personnel and equipment delivered with a given vehicle number. Ant colony system with maximum-minimum limit is adopted to solve these problems. The distance of each route is computed and the route with maximum distance is chosen. The objective is to minimize the maximum distance. The algorithm is implemented and tested on some instances. The results demonstrate the effectiveness of the method.

With the rapid growth in application of series data mining, one important issue is discovering character patterns in larger data sets. Two limitations of previous works were the weak efficiency and rigid partition. In this paper, we introduce a novel pattern searching algorithm that using cloud models to implement concept hierarchies and data reduction. The reduction in this algorithm is based on symbolic mapping which uses cloud transformation method. Compared with other works, we make use of linguistic atoms to describe series character both specifically and holistically. Furthermore, being the fuzzy and probabilistic of cloud models, soft partition to continuous numeric attributes and the capability to data noise were realized. Normal segmentation method was done as comparison to show the performance of cloud models based algorithm. The efficiency is improved obviously. Moreover, noise-adding experiment was implemented to show that algorithm has robustness to the noise.

This paper introduces the theory of super-resolution image reconstruction and degraded model in brief, and presents a new super-resolution image reconstruction algorithm .The algorithm bases on the new image registration excluded aliased frequency domain and the Projection Onto Convex Set (POCS) method. The algorithm can precisely estimate the image registration parameter by excluding aliased frequency domain of the low-resolution images and killing the center part of the magnitude spectrum. In order to compute the shifts and the rotation angle, we set up the polar coordinates in the center of the image. By computing the frequency function of the rotation angle by integrating over radial lines, the algorithm converts the two-dimension correlation to one-dimension correlation. And then, the POCS method is used to reconstruct high-resolution image from these aliased image sequences. As a result, we find that the reconstruction algorithm has the same precision of image registration as the spatial image registration and good effect of super-resolution image reconstruction.

The cooperative optimization is a newly discovered metaheuristic for solving difficult combinatorial optimization problems. It is inspired by the cooperation principle in social systems where individuals in a system often work together in a cooperative way to solve hard problems of a complexity beyond the capability of any individual in the system. Unlike any existing metaheuristics, it has a number of global optimality conditions so that the cooperative optimization algorithms know where to find global optima and when to stop searching. Furthermore, a cooperative optimization algorithm has a unique equilibrium and converges to it with an exponential rate regardless of initial conditions and perturbations. In solving real-world optimization problems, the cooperative optimization algorithms have often significantly outperformed state-of-the-art algorithms.

This paper introduces an unsupervised adaptive principal components analysis (APEX) neural network (NN) for blind pseudo noise (PN) sequence extraction of lower signal to noise ratios (SNR) direct sequence spread spectrum (DS-SS) signals. The proposed method is based on eigen-analysis of DS-SS signals. As the eigen-analysis method is based on the decomposition of autocorrelation matrix of signals, it has computational defects when the signal vectors became longer, etc. So, we introduce the APEX NN to extract the PN sequence blindly. We also make complexity analysis of the proposed method and comparison with the other methods. Theoretical analysis and computer simulations verify the effectiveness of the method.

Data gathering is one of the most important processes in wireless sensor networks (WSN). The sensor nodes gather information and send it to a base station, which consumes significant amounts of power. Due to the limited battery life, energy efficiency is becoming a major challenging problem in WSN. Some energy- efficient data gathering protocols are proposed for WSN such as LEACH, PEGASIS, and PEDAP. But these protocols still have some disadvantages and are not fit for our application, wireless sensor transportation monitoring network (WSTMN). So we present a new chain-based data gathering protocol, CBDGP. The minimum total energy algorithm (MTEA) is used to construct the chain. Moreover, the layered minimum total energy chain construction algorithm with delay reducing is presented for CBDGP. The results show that CBDGP works for WSTMN better than LEACH and PEGASIS. It prolongs the lifetime and reduces the delay of WSTMN.

In this paper, we develop an energy and bandwidth efficient approach for target classification in sensor networks. Instead of adopting decision fusion to reduce network traffic as some recent research, we try to realize energy efficient target classification from a computational point of view. Our contribution is we propose a novel tree construction algorithm that autonomously organizes the distributed computation resources to execute the trained BP-network (BPN) in parallel manner. We evaluate the performance of our parallel computing paradigm compared to the traditional client/server-based computing paradigm from perspectives of energy consumption and communication traffic through analytical study. Finally, we take a target classification experiment to show the effectiveness of the proposed computing paradigm.

A mobile ad hoc network (MANET) is an autonomous system of mobile nodes connected by wireless links. There is no static infrastructure such as base station in cell mobile communication. Due to the dynamic nature of the network topology and restricted resources, quality of service (QoS) and multicast routing in MANET is a challenging task. Finding and maintaining QoS multicast routing in the data is still more challenging. In this paper, we present an entropy-based genetic algorithm (GA) to support QoS multicast routing in mobile ad hoc networks (EQMGA). The key idea of EQMGA algorithm is to construct the new metric-entropy and select the long-life path with the help of entropy metric to reduce the number of route reconstruction so as to provide QoS guarantee in the ad hoc network. The simulation results demonstrate that the proposed approach and parameters provide an accurate and efficient method to estimate and evaluate the route stability in dynamic mobile networks.

The focus of situation-aware ubiquitous computing has increased lately. An example of situation-aware applications is a multimedia education system. The development of multimedia computers and communication techniques has made it possible for a mind to be transmitted from a teacher to a student in distance environment. This paper proposes an Adaptive Fault Tolerance (AFT) algorithm in situation-aware middleware framework and presents its simulation model of AFT-based agents. FTE(Fault Tolerance Environment) provide several functions and features capable of developing multimedia distant education system among students and teachers during lecture. AFT is a system that is suitable for detecting and recovering software error based on distributed multimedia education environment as FTE by using software techniques. This method detects an error by using process database. When an error occurs, FTA(Fault Tolerance Agent) inspects it by using API(Application Program Interface) function for process database. If an error is found, FTA decides whether it is hardware error or software error. In case of software error, it can be recoverable. The purpose of AFT system is to maintain and recover for FTE session automatically. This paper proposes an Adaptive Fault Tolerance (AFT) algorithm in situation-aware middleware framework and presents its simulation model of AFT-based agents.

This paper proposes a new model of access control by analyzing the window and attributes of the object, and based on this, a mechanism that offers a seamless multimedia view without interfering with access control is also suggested. There are two approaches to software architecture on which applications for multimedia distance education environment in situation-aware middleware are based. Those include CACV(Centralized-Abstraction and Centralized-View) and RARV(Replicated-Abstraction and Replicated-View).To win over such dilemma for centralized or replicated architecture, a combined approach, CARV(the Centralized Abstraction and Replicated View) architecture is used to realize the application sharing agent.

Wireless Sensor Network is an advanced technology that has a variety of applications such as environmental monitoring, battlefield, medical system, and crop precision. Minimizing power consumption of node is important in these networks. Transmit power control and clustering can reduce the energy consumption efficiently when nodes are non-homogeneously dispersed in space. This paper presents the clustering algorithm in wireless sensor networks. The clustering algorithm is based on the optimization of transmit power level by using the soft computing approaches. This solution determines the node transmit power level statistically and achieves energy savings efficiently.

In this paper, a method of controlling packet transmission rate between nodes on an Ad-hoc network is proposed considering the characteristics of IEEE 802.11 possessing different transmission efficiencies by different transmission distances. There have been a lot of researches about algorithms for efficient routing and power saving using static power sources in the field of mobile Ad-hoc networks up until now. However, those researches have been conducted only on the assumption of ideal experimental cases. This paper considers the way of finding adequate transmission rate for the transmission distances between nodes on a mobile Ad-hoc networks so that a more realizable method is presented. In this research, a controlling algorithm for transmission data rates by the distances between mobile nodes is realized using Fuzzy logic, possibly available to be applied to Ad-hoc network routing.

Time synchronization in wireless sensor networks is critical for providing accurate timing service to many applications. This paper presents an Error Statistic based Time Synchronization protocol (ESTS) for wireless sensor networks. ESTS uses a flooding mechanism for basic global time synchronization and delay measurement, and achieves a fine grained synchronization through compensating per hop delay to all the nodes. It is a light weight mechanism for its simplicity in both computation and communication. Experiments show that ESTS obtains a better performance in energy efficiency than most in-situ time synchronization protocols in the same precision level.

This paper presents a new algorithm, Extra, for extending the lifetime of ad hoc wireless networks. Extra tries to conserve energy by identifying and switching off nodes that are momentarily redundant for message routing in the network. Extra is independent of the underlying routing protocol and uses exclusively information that is collected locally. Simulation studies conducted have shown promising results.

The paper entitled “Recognition of Emotion with SVMs” starting on page 701 of this volume has been withdrawn due to a serious case of plagiarism.