Advances in Computation and Intelligence

In e-learning environment, online testing system can help to evaluate students’ learning status precisely. To meet the users’ multiple assessment requirements, a new test-sheet composition model was put forward. Based on the proposed model, a genetic algorithm with effective coding strategy and problem characteristic mutation operation were designed to generate high quality test-sheet in online testing systems. The proposed algorithm was tested using a series of item banks with different scales. Superiority of the proposed algorithm is demonstrated by comparing it with the genetic algorithm with binary coding strategy.

It is a often hard to find a feasible and optimum route because the routing resources are constrained in coarse-grain RA (CGRA) and several functions are often defined in same one element of RA. In this paper, a proposed propriety-based path encoding genetic algorithm is applied for the routing problem of CGRA. By mapping Fast Fourier Transform of the butterfly computation to CTaiJi that is a newly developed CGRA, the proposed GA shows good ability to find the good solution.

This paper employs a genetic algorithm to evolve an optimized stock market prediction system. The prediction based on a range of technical indicators generates signals to indicate the price movement. The performance of the system is analyzed and compared to market movements as represented by its index. Also investment funds run by professional traders are selected to establish a relative measure of success. The results show that the evolved system outperforms the index and funds in different market environments.

Genetic Algorithm (GA) is a powerful tool for science computing, while Parallel Genetic Algorithm (PGA) further promotes the performance of computing. However, the traditional parallel computing environment is very difficult to set up, much less the price. This gives rise to the appearance of moving dense computing to graphics hardware, which is inexpensive and more powerful. The paper presents a hierarchical parallel genetic algorithm, implemented by NVIDIA’s Compute Unified Device Architecture (CUDA). Mixed with master-slave parallelization method and multiple-demes parallelization method, this algorithm has contributed to better utilization of threads and high-speed shared memory in CUDA.

Identification of transcription factor binding sites (TFBSs) or motifs plays an important role in deciphering the mechanisms of gene regulation. Although many experimental and computational methods have been developed, finding TFBSs remains a challenging problem. We propose and develop a novel sampling based motif finding method coupled with PSFM optimization by genetic algorithm, which we call Motif GibbsGA. One significant feature of Motif GibbsGA is the combination of Gibbs sampling and PSFM optimization by genetic algorithm. Based on position-specific frequency matrix (PSFM) motif model, a greedy strategy for choosing the initial parameters of PSFM is employed. Then a Gibbs sampler is built with respect to PSFM model. During the sampling process, PSFM is improved via a genetic algorithm. A post-processing with adaptive adding and removing is used to handle general cases with arbitrary numbers of instances per sequence. We test our method on the benchmark dataset compiled by Tompa et al. for assessing computational tools that predict TFBSs. The performance of Motif GibbsGA on the data set compares well to, and in many cases exceeds, the performance of existing tools. This is in part attributed to the significant role played by the genetic algorithm which has improved PSFM.

The medical waste management is of great importance because of the potential environmental hazards and public health risks. Manufacturers have to collect the medical waste and control its recovery or disposal. Medical waste recovery, which encompasses reusing, remanufacturing and materials recycling, requires a specially structured reverse logistics network in order to collect the medical waste efficiently. This paper presents a Mixed Integer Linear Programming model with minimizing costs for medical waste reverse logistics networks. The total costs for reverse logistics include transportation cost, fixed cost of opening the collecting centers and processing centers and operation cost at these facilities over finite planning horizons. An improved genetic algorithm method with a hybrid encoding rule is used to solve the proposed model. The efficiency and practicability of the proposed model is validated by an application to an illustrative example dealing with medical waste returned from some hospitals to a given manufacture.

In this paper, a new algorithm named SGEGC was proposed. Inspired by selfish gene theory, SGEGC uses a vector of survival rate to model the condition distribution, which serves as a virtual population that is used to generate new individuals. While the present Estimation of Distribution Algorithms (EDAs) require much time to learn the complex relationships among variables, SGEGC employs an approach that exchanges the relevant genetic components. Experimental results show that the proposed approach is more efficient in convergent reliability and convergent velocity in comparison with BMDA, COMIT and MIMIC in the test functions.

This paper presents a heuristic fast-matching algorithm for quantum reversible logic circuits synthesis. The algorithm uses the quantum gate formula as the heuristic rules to perform forward-matching. This can avoid blind matching and decrease the matching complexity. Simple, smaller overhead and better to simplify circuits, the algorithm can perform well in the multi-quantum reversible logic circuits synthesis.

Reversible logic studies have promising potential on energy lossless circuit design, quantum computation, nanotechnology, etc. This paper proposes an analogic selection sorting algorithm essentially based on the transformation-based algorithm. It uses an unweighted, undirected graph for the representation of all transformable paths. During the synthesis process, a sequence of transformations are performed to enable all the output patterns to appear in the right place. The whole process can be implemented by a sequence of Toffoli gates. In addition, a simplification algorithm is put forward to further optimize the generated circuit. The experimental results show that this algorithm, compared with other exact methods, can achieve optimal or nearly optimal solutions with less computation time. Furthermore, it is more easily understood and implemented.

The machine-part cell formation is a NP- complete combinational optimization in cellular manufacturing system. Previous researches have revealed that although the genetic algorithm (GA) can get high quality solutions, special selection strategy, crossover and mutation operators as well as the parameters must be defined previously to solve the problem efficiently and flexibly. The Estimation of Distribution Algorithms (EDAs) has recently been recognized as a new computing paradigm in evolutionary computation which can overcome some drawbacks of the traditional GA mentioned above. In this paper, two kinds of the EDAs, UMDA and EBNA

BIC

are applied to solve the machine-part cell formation problem. Simulation results on six well known problems show that the UMDA and EBNA

BIC

can attain satisfied solutions more simply and efficiently.

This paper investigates global convergence for a novel class of delayed neural networks with non-Lipschitz neuron activations and impulses based on the topological degree theory and Lyapunov functional method. Some suffcient conditions are derived to ensure the existence, and global exponential stability of the equilibrium point of neural networks. Finally, a numerical example is given to demonstrate the effectiveness of the obtained result.

In this paper, by following Zhang

et al

’s neural-dynamic method proposed formally since March 2001, two recurrent neural networks are generalized to solve online the time-varying convex quadratic-minimization and quadratic-programming (QP) problems, of which the latter is subject to a time-varying linear-equality constraint as an example. In comparison with conventional gradient-based neural networks or gradient neural networks (GNN), the resultant Zhang neural networks (ZNN) can be unified as a superior approach for solving online the time-varying quadratic problems. For the purpose of time-varying quadratic-problems solving, this paper investigates comparatively both ZNN and GNN solvers, and then their unified modeling techniques. The modeling results substantiate well the efficacy of such ZNN models on solving online the time-varying convex QP problems.

Implementing target cost control at the design stage can better reduce the cost. However, for automakers in the Chinese market, no adequate attention is paid to the target cost control during design stage for various reasons. Among these reasons, the lack of an effective cost control tool is a substantial one. In this study, a target cost control method is proposed and artificial intelligence is employed for new product cost reduction. At the early design stage, Back Propagation (BP) neural network is introduced to estimate and evaluate the target cost of different designs. Consequently, a cost saving design can be chosen. The target cost can be mainly achieved through procurement cost control. A procurement model is designed for balancing procurement cost reduction and supplier satisfaction. To search the optimal solution for this model, genetic algorithm is introduced. A case study of the proposed method in a Chinese automobile company is also discussed .

This paper first gives a general review of the present situation of the hard real time power saving scheduling algorithm and then demonstrates its currently related technology and several classic power saving scheduling algorithms. Moreover, it studies the power saving sheduling algorithm of hard real time periodical task based on DVS technology and sonducts a comparison simulation of three different scheduling algorithms.

Based on the Chen-Harker-Kanzow-Smale smoothing function, a smoothing Newton method is developed for solving the symmetric conic linear programming. Without any restrictions for its starting point, this algorithm solves only one linear system of equations at each iteration and proves to be globally convergent in absence of uniform nonsingularity. Numerical results indicate that it is promising in future applications.

Particle Swarm Optimization (PSO) algorithm was developed under the inspiration of behavior laws of bird flocks, fish schools and human communities. In order to get rid of the disadvantages of standard Particle Swarm Optimization algorithm like being trapped easily into a local optimum, this paper improves the standard PSO and proposes a new algorithm to solve these problems. The new algorithm keeps not only the fast convergence speed characteristic of PSO, but effectively improves the capability of global searching as well. Compared with standard PSO on the Benchmarks function, the new algorithm produces more efficient results.

Clustering is a core problem in data mining and machine learning though it is widely applied in many fields. Recently, it is very popular to use the evolutionary algorithm to solve the problem. This paper proposes an automatic clustering differential evolution (DE) technique for the problem. This approach can be characterized by (i) proposing a modified point symmetry-based cluster validity index (CVI) as a measure of the validity of the corresponding partitioning, (ii) using the Kd-tree based nearest neighbor search to reduce the complexity of finding the closest symmetric point, and (iii) employing a new representation to represent an individual. Experiments conducted on 6 artificial data sets of diverse complexities indicate that this approach is suitable for both the symmetrical intra-clusters and the symmetrical inter-clusters. In addition, it is able to find the optimal number of clusters of the data. Furthermore, based on the comparison with the original point symmetry-based CVI, this proposed point symmetry-based CVI shows better performance in terms of the F-measure and the number of clusters found.

By introducing an adaptive error function, a balanced ensemble learning had been developed from negative correlation learning. In this paper, balanced ensemble learning had been used to train a set of small neural networks with one hidden node only. The experimental results suggest that balanced ensemble learning is able to create a strong ensemble by combining a set of weak learners. Different to bagging and boosting where learners are trained on randomly re-sampled data from the original set of patterns, learners could be trained on all available data in balanced ensemble learning. It is interesting to be discovered that learners by balanced ensemble learning could be just be slightly better than random guessing even if they had been trained on the whole data set. Another difference among these ensemble learning methods is that learners are trained simultaneously in balanced ensemble learning when learners are trained independently in bagging, and sequentially in boosting.

Variogram is a basic tool of geostatistics, used to describe the randomicity and structural property of regionalized variable. While estimating variogram parameters is the basic issue of spatial statistics analysis. Estimate of the parameters is always made by using the theoretical variogram model to fit experimental variogram model. However, it is difficult to obtain the optimization results because the theoretical variogram is not successively derivable if traditional numerical algorithm is used. Differential evolution algorithm is a new evolution algorithm which adopts real number encoding format and has a fast convergence. In this paper, it is the first time to use differential evolution algorithm to estimate variogram parameters. Orthogonal experiments are conducted to ensure the diversity of initial species. The results illustrate that the approach of DE can work out the problem with fast convergence, strong optimization and excellent stability.

Some types of space trajectory design problems present highly multimodal, globally non-convex objective functions with a large number of local minima, often nested. This paper proposes some memetic strategies to improve the performance of the basic heuristic of differential evolution when applied to the solution of global trajectory optimisation. In particular, it is often more useful to find families of good solutions rather than a single, globally optimal one. A rigorous testing procedure is introduced to measure the performance of a global optimisation algorithm. The memetic strategies are tested on a standard set of difficult trajectory optimisation problems.

With a comparison to the random selection approach used in the weight-based multiobjective immune algorithm (WBMOIA), this paper proposes a new selection approach based on the truncation algorithm with similar individuals (TASI). Then the effect of the proposed selection approach is examined on the performance of WBMOIA. On one hand, the performance is compared between WBMOIA with the random selection approach and WBMOIA with the proposed selection approach. On the other hand, simulation results on a number of problems are presented to investigate if there exists any value of the reduction rate where WBMOIA performs well. Experiment results show that the performance of WBMOIA can be improved by the proposed selection approach and a better reduction rate can be obtained for each test problem.

This paper presents a novel hybrid evolutionary algorithm for function optimization. In this algorithm, the space transformation search (STS) is embedded into a novel genetic algorithm (GA) which employs a novel crossover operator based on a nonconvex linear combination of multiple parents and elite-preservation strategy (EGT). STS transforms the search space to increase more opportunities for finding the global optimum and accelerate convergence speed. Experimental studies on 15 benchmark functions show that the STS-EGT not only has good ability to help EGT jump out of local optimum but also obtains faster convergence than the STS-GT which has no elitepreservation strategy.

Kalman filter perform optimally when the noise statistics for the measurement and process are completely known in integrated navigation system. However, the noise statistics could change with the actual working environment and so the initial priori value would represent the actual state of noise incorrectly. To solve this problem, this paper presents an adaptive Kalman Filter based on evolutionary algorithm. The hybrid method improves the real-time noise statistics by the procedure of global search. Field test data are processed to evaluate the performance of the proposed method. The results of experiment show the proposed method is capable of improving the output precision and adaptive capacity of filtering, and thus is valuable in application.

Many real-life problems can be formulated as numerical optimization of certain objective functions. However, for an objective function possesses numerous local optima, many evolutionary algorithms (EAs) would be trapped in local solutions. To improve the search efficiency, this paper presents a clone and Cauchy-mutation evolutionary algorithm (CCEA), which employs dynamic clone and Cauchy mutation methods, for numerical optimization. For a suit of 23 benchmark test functions, CCEA is able to locate the near-optimal solutions for almost 23 test functions with relatively small variance. Especially, for

f

14

-

f

23

, CCEA can get better solutions than other algorithms.

This paper is a technical supplement of our original research in the combination of genetic programming (GP), Hoare logic, model checking, and finite state automaton for Hoare triple constructions. Although there is no problem in achieving this goal by first constructing the generalized models for some given Hoare triples, the desired Hoare triples from the application of general GP approaches to them still produce ineffective results and hence needs further improvement. In this paper, we solve it through the use of some semantically valid genetic operations. Precisely, we check logic relationships among Hoare triples and generalized models for their consistence.

In order to ensure the target selection of the cobalt-rich crust exploration and thus to benefit the mining area application in the future, it is necessary to conduct the method research for the resources calculation and the exploration area circling. In this paper, we employ the fractal parameters of the seamount terrain isoline to explore the relationship, that exists in the distribution and enrichment of the seamount cobalt-rich crusts and the seamount distribution fractal parameters, to provide quantitative basis for judgment of the blocks to be estimated. Finally, we carry out an empirical research with the terrain and station sampling data of a certain seamount in the Central Pacific and establish the scheme of circle mines by selecting the suitable threshold value of fractal box dimension. Then we use Ordinary Kriging method to complete the resources evaluation of cobalt-rich crust.

This paper introduces a hybrid algorithm called as the HENSA-SAT for the large-scale Satisfiability (SAT) problems. The HENSA-SAT is the hybrid of Evolutionary Negative Selection Algorithm (ENSA), the Flip Heuristic, the

BackForwardFlipHeuristic

procedure and the

VerticalClimbing

procedure. The Negative Selection (NS) is called twice for different purposes. One is used to make the search start in as many different areas as possible. The other is used to restrict the times of calling the

BackForwardFlipHeuristic

for local search. The Flip Heuristic, the

BackForwardFlipHeuristic

procedure and the

VerticalClimbing

procedure are used to enhance the local search. Experiment results show that the proposed algorithm is competitive with the GASAT that is the state-of-the-art algorithm for the large-scale SAT problems.

Recently, Evolutionary Algorithms (EAs) with associative memory schemes have been developed to solve Dynamic Optimization Problems (DOPs). Current associative memory schemes always retrieve both the best memory individual and the corresponding environmental information. However, the memory individual with the best fitness could not be the most appropriate one for new environments. In this paper, two novel associative memory retrieving strategies are proposed to obtain the most appropriate memory environmental information. In these strategies, two best individuals are first selected from the two best memory individuals and the current best individual. Then, their corresponding environmental information is evaluated according to either the survivability or the diversity, one of which is retrieved. In experiments, the proposed two strategies were embedded into the state-of-the-art algorithm, i.e. the MPBIL, and tested on three dynamic functions in cyclic environments. Experiment results demonstrate that the proposed retrieving strategies enhance the search ability in cyclic environments.

The relevance of traditional classification methods, such as CBA and CMAR, bring the problems of frequent scanning the database, resulting in excessive candidate sets, as well as the complex construction of FP-tree that causes excessive consumption. This paper studies the classification rules based on association rules - MCAR (Mining Class Association Rules). The database only needs scanning once, and the cross-support operation is used for the calculation as the format of databases is vertical layout for easily computing the support of the frequent items. Not only the minimum support and minimum confidence is used to prune the candidate set, but also the concept of class-frequent items is taken into account to delete the rules that may hinder the effective improvement of the algorithm performance.

Classification algorithms as an important technology in data mining and machine learning have been widely studied and applied. Many methods can be used to build classifiers, such as the decision tree, Bayesian method, instance-based learning, artificial neural network and support vector machine. This paper focuses on the classification methods based on decision tree learning, Bayesian learning, and instance-based learning. In each kind of classification methods, many improvements have been presented to scale up the classification accuracy of the basic algorithm. The paper also studies and compares the classification performance on classification accuracy empirically, using the whole 36 UCI data sets obtained from various sources selected by Weka. The experiment results re-demonstrate the efficiency of all these improved algorithms.

The majority of machine learning algorithms previously designed usually assume that their training sets are well-balanced, but data in the real-world is usually imbalanced. The class imbalance problem is pervasive and ubiquitous, causing trouble to a large segment of the data mining community. As the conventional machine learning algorithms have bad performance when they learn from imbalanced data sets, it is necessary to find solutions to machine learning on imbalanced data sets. This paper presents a novel Isomap-based hybrid re-sampling approach to improve the conventional SMOTE algorithm by incorporating the Isometric feature mapping algorithm (Isomap). Experiment results demonstrate that this hybrid re-sampling algorithm attains a performance superior to that of the re-sampling. It is clear that the Isomap method is an effective means to reduce the dimension of the re-sampling. This provides a new possible solution for dealing with the IDS classification.

Intrusion detection has been extensively studied in the last two decades. However, most existing intrusion detection techniques detect limited number of attack types and report a huge number of false alarms. The hybrid approach has been proposed recently to improve the performance of intrusion detection systems (IDSs). A big challenge for constructing such a multi-sensor based IDS is how to make accurate inferences that minimize the number of false alerts and maximize the detection accuracy, thus releasing the security operator from the burden of high volume of conflicting event reports. We address this issue and propose a hybrid framework to achieve an optimal performance for detecting network traffic anomalies. In particular, we apply SNORT as the signature based intrusion detector and the other two anomaly detection methods, namely non-parametric CUmulative SUM (CUSUM) and EM based clustering, as the anomaly detector. The experimental evaluation with the 1999 DARPA intrusion detection evaluation dataset shows that our approach successfully detects a large portion of the attacks missed by SNORT while also reducing the false alarm rate.

Economic models for mineral resources assessment are transferring from single objective to multiple objectives nowadays. However, common approaches to solve these multi-criteria problems are still staying in single-objective methods, by combining all objective functions into a single functional form, but such methods can only obtain one solution. In this paper, NSGA-II,a multiobjective optimization evolutionary algorithm, is adopted to optimize multiple objectives of mineral resource exploitation.Two case study prove that NSGA-II can offer multiple solutions and be irrelevant with starting point, moreover, results by NSGA-II are better than references.

This paper proposes a robust and efficient eye state detection method based on an improved algorithm called LBP+SVM mode. LBP (local binary pattern) methodology is first used to select the two groups of candidates from a whole face image. Then corresponding SVMs (supporting vector machine) are employed to verify the real eye and its state. The LBP methodology makes it robust against rotation, illumination and occlusion to find the candidates, and the SVM helps to make the final verification correct.

A neural network architecture is introduced for context-sensitive binding processing in visual cortex areas such as perceptual grouping, attention and boundary-surface interaction. The present architecture, based on LAMINART and FAÇADE theory, shows how layered circuits in cortex areas enable feedforward, horizontal, and feedback interactions and intractions, together with balanced connections, to complete perceptual groupings with attention modulation. Thus, a pre-attentive/attentive interface for cortex has been established within the same circuits. Moreover, this architecture exhibits the invisible perceptual grouping and visible surface filling-in interaction with complementary computing property, introduced in FAÇADE theory. By implementing SOC filtering, preattentive perception can response well to contrast-sensitive stimuli. Also the simulations illustrates selective propagation of the attention along an object grouping as well as surface filling and the protection of them from competitive masking. At the same time it demonstrates the generation of attention modulation by boundary-surface interactions as well as translation of the attention across complementary visual processing streams.

Neighborhood selection plays an important role in manifold learning algorithm. This paper proposes an Adaptive Neighborhood Select algorithm based on Local Linearity(ANSLL). Given manifold dimensionality

d

as a priori knowledge, ANSLL algorithm constructs neighborhood based on two principles: 1. data points in the same neighborhood should approximately lie on a

d

-dimensional linear subspace; 2. each neighborhood should be as large as possible. And in ASNLL algorithm PCA technique is exploited to measure the linearity of finite data points set. Moreover, we present an improved method of constructing neighborhood graph, which can improve the accuracy of geodesic distance estimate for isometric embedding. Experiments on both synthetic data sets and real data sets show that ANSLL algorithm can adaptively construct neighborhood according to local curvature of data manifold and then improve the performance of most manifold algorithms, such as ISOMAP and LLE.

Recently, spam has become an increasingly important problem. In this paper, a support vector machine (SVM) is used as the spam filter. Then a study is made of the effect of classification error rate when different subsets of corpora are used, and of the filter accuracy when SVM’s with linear, polynomial, or RBF kernels is used. Also an investigation is made of the effect of the size of attribute sets. Based on the experimental results and analysis, it is concluded that SVM will be a very good alternative for building anti-spam classifiers, with consideration of a good combination of accuracy, consistency, and speed.

The reasonalbe and effective determination of the weight allocation is very critical to multi-attribute decision-making. This paper presents a novel multi-attribute weight allocation method based on the fuzzy clustering analysis and the information entropy theory in rough sets theory. It first studies the fuzzy clustering analysis method based on fuzzy transitive closure with the introduction of the information entropy theory in rough sets. Furthermore, it discusses the detailed steps of the proposed approach thoroughly . After the fuzzy clustering of the source data, the overall reasonable threshold is extracted based on F-statistics and the multi-attribute weight allocation is obtained using the information entropy theory. Finally, a case study is given to show the reasonability and validity of the proposed approach.

This paper focuses on spatio-temporal non-linear intelligent prediction modeling for regional data, and discusses the application of Back-Propagation neural network (BPN) into analysis of regional data in geographic information system (GIS). With their characteristics of space-dependence and space volatility, the regional data determine the accuracy of the prediction model.With consideration of the sectional instability of the spatial pattern, the paper brings forward a modeling method based on regional neural network. First, the space units of researching regions are divided into different sub-regions by improved K-means algorithm based on spatial adjacency relationship, corresponding with sub-regions. Then, a modular BP network is built up, which is composed with main network, gate-network and sub-network. This network is thus named as regional spatio-temporal neural network (RSTN) model. Afterwards, the sub-networks are traiend respectively for every sub-region, and the output of sub-networks is input of main network with adjustment of gate-network The output of main network is predictive results. The spatio-temporal predictive capability of model is measured by average variance rate (AVR) and dynamic similar rate (DSR). At last, the RSTN model and the global BPN model are compared by the analysis of an example: prediction for influenza cases of 94 countries in France. The comparison declares that RSTN model has more powerful prediction capability.

The present paper introduces the development, valuable part and application of neural network. It also analyzes systematically the existing problems and the combination of neural network with wavelet analysis, fuzzy set, chaos, rough sets and other theories, together with its applications and the hot spots of the research on neural network. The analysis proves that the prospects of neural network will be primising with the combination method, and that subject integration will be the chief interest for the neural network research.

An approach to solve complex combinatorial optimizations problems is the Ant Colony Optimization Metaheuristic (ACO). There are several variations of this metaheuristic. One of them, the Max-Min Ant System, is an algorithm that presents excellent performance for some classes of combinatorial problems, such as Traveling Salesman Problem and the Quadratic Assignment Problem. This paper presents a method of convergence control of the Max-Min variation of Ant Colony Optimization Metaheuristic using paraconsistent logic. The proposed method can be adapted to any variation of the Ant Colony Optimization Metaheuristic.

An Artificial Neural Network (ANN) is an information processing paradigm inspired by the biological nervous systems. It is composed of a large number of highly interconnected processing elements (neurones) working in unison to solve specific problems. The negative correlation learning encourages different individual network to study and trains different parts of the ensemble in order to make the whole ensemble study the whole training data better. This paper improves the method of negative correlation learning by using a BP algorithm with impulse in the error function. The method is an algorithm in batches with more powerful generalization and study speed because it combines primitive correlation learning with BP algorithm of impulse.

Floorplanning is a critical step in the physical design of Very Large Scale Integrated (VLSI) circuits. Its main target is optimizing the layout area and interconnection wire length of chips, which can be transformed into a Multi-objective Optimization Problem (MOP). In this paper, we propose a discrete Particle Swarm Optimization (PSO) algorithm for MOP which could take many key objectives into consideration and give a good compromise between them. The experiments on MCNC benchmarks show that the proposed algorithm is effective, and gives out many optional results for user’s choice according to partialness, which can not be finished by traditional methods.

An improved particle swarm optimization algorithm, CSPSO (Chaotic Species-based particle swarm optimization), is proposed for solving the template matching problem. Template matching is one of the image comparison techniques widely applied to component existence checking in the printed circuit board (PCB) and electronics assembly industries. The proposed approach adopts the special nonlinear characteristic and ergodicity of chaos to enrich the search ability of the species-based particle swarm optimization (SPSO). To test its performance, the proposed CSPSO-based approach is compared with SPSO-based approach using two experimental studies. The CSPSO-based approach is proven to be superior to the original SPSO-based one in term of efficiency.

Many optimization problems in real world are dynamic in the sense that the global optimum value and the shape of fitness function may change with time. The task for the optimization algorithm in these environments is to find global optima quickly after the change in environment is detected. In this paper, we propose a new hybrid model of particle swarm optimization and cellular automata which addresses this issue. The main idea behind our approach is to utilized local interactions in cellular automata and split the population of particles into different groups across cells of cellular automata. Each group tries to find an optimum locally which results in finding the global optima. Experimental results show that cellular PSO outperforms mQSO, a well known PSO model in literature, both in accuracy and complexity in a dynamic environment where peaks change in width and height quickly or there are many peaks.

The layout design with dynamic performance constraints belong to NP-hard problem in mathematics, optimized with the general particle swarm optimization (PSO), to slow down convergence and easy trap in local optima. This paper, taking the layout problem of satellite cabins as background, proposed an adaptive particle swarm optimizer with a excellent search performance, which employs a dynamic inertia factor, a dynamic graph planeradius and a set of dynamic search operator of space and velocity, to plan large-scale space global search and refined local search as a whole in optimization process, and to quicken convergence speed, avoid premature problem, economize computational expenses, and obtain global optimum. The experiment on the proposed algorithm and its comparison with other published methods on constrained layout examples demonstrate that the revised algorithm is feasible and efficient.

Many real-world problems are dynamic, requiring an optimization algorithm which is able to continuously track a changing optimum over time. This paper presents a new variant of Particle Swarm Optimization (PSO) specifically designed to work well in dynamic environments. The main idea is to divide the population of particles into a set of interacting swarms. These swarms interact locally by dynamic regrouping and dispersing. Cauchy mutation is applied to the global best particle when the swarm detects the environment of the change. The dynamic function (proposed by Morrison and De Jong) is used to test the performance of the proposed algorithm. The comparison of the numerical experimental results with those of other variant PSO illustrates that the proposed algorithm is an excellent alternative to track dynamically changing optima.

Three-dimensional (3-D) braided composite, a new type composite, has the better performance designable characteristic. This paper describes the damping analysis of the hollow-rectangular-section three-dimensional braided composite made by 2-step method with a consideration of the wide application of hollow-rectangular-section three-dimensional braided composite in engineering. Then, it proposes the mathematical models for optimization of the damping of the three-dimensional braided composite. The objective functions are based on the specific damping capacity of the composite, and the design variables are the braiding parameters and sectional geometrical size of the composite. The results of numeral examples show that the better damping characteristic could be obtained by using optimal design with particle swarm optimization (PSO) algorithm, contenting the determinate restriction. The method proposed here is useful for the design and engineering application of the kind of member.

A novel parallel hybrid particle swarm optimization algorithm named hmPSO is presented. The new algorithm combines particle swarm optimization (PSO) with a local search method which aims to accelerate the rate of convergence. The PSO provides initial guesses to the local search method and the local search accelerates PSO with its solutions. The hybrid global optimization algorithm adjusts its searching space through the local search results. Parallelization is based on the client-server model, which is ideal for asynchronous distributed computations. The server, the center of data exchange, manages requests and coordinates the time-consuming objective function computations undertaken by individual clients which locate in separate processors. A case study in geotechnical engineering demonstrates the effectiveness and efficiency of the proposed algorithm.

Storage-based intrusion detection systems (SIDS) allow storage systems to watch for suspicious activity. This paper presents a novel storage- based intrusion detection scheme to monitor the user’s activities with the artificial immune technique. Compared with the previous SIDS prototype, the SIDS using artificial immune technique can recognize a strange suspicious behavior. Before simulation, a set of appropriate parameters of algorithm are fitted according to the mean convergence speed and detection efficiency. The simulation shows the proposed scheme can reach higher detection rate and lower false alarm rate than the previous ones.

This paper put forward a new method for Optimal Configuration of Weapon System (OCWS). It combines PROMETHEE II with Correspond analysis and projects the decision space on the two-dimensional principal component plane, in which R-type principal component refers to each attribute, and Q-type principal component means each alternative. This makes it possible to show the properties, such as comparability or otherness between different alternatives, the importance or conflict between different attributes and the sensitivity of compromised solution, more simply and conveniently, intuitionally. Therefore, it proves an excellent visual method for OCWS.

Mobile ad hoc networking (MANET) bring great challenges in security due to its high dynamics, link vulnerability and complete decentralization. With routing being a critical aspect for MANETs, existing routing protocols however are not sufficient for security requirements. In this paper, we present a route discovery algorithm that mitigates the detrimental effects of malicious behavior, as to provide correct connectivity information. Our algorithm guarantees that fabricated, compromised, or replayed route replies would either be rejected or never reach back the querying node. Furthermore, the algorithm responsiveness is safeguarded under different types of attacks that exploit the routing algorithm itself. The sole requirement of the proposed scheme is the existence of a security association between the node initiating the query and the sought destination. Specifically, no assumption is made regarding the intermediate nodes, which may exhibit arbitrary and malicious behavior. The scheme is robust in the presence of a number of non-colluding nodes, and provides accurate routing information in a timely manner.

Opening property of P2P network allows nodes to freely join P2P network and to create identity at no cost. Utilizing the loopholes, malicious nodes can create a number of identities in a short time, which will exhaust the identifiers resources and damage the operation of P2P network. The paper proposes a conundrum verification scheme which enables the node to join P2P network and creates identity more difficult. Moreover, it also puts forward a detection and elimination scheme, which can help P2P network to detect identity attackers promptly and eliminate them. Simulation experiments demonstrate that with the combination of the two schemes, P2P network can prevent identity attack effectively.

License Plate Recognition is an important component of Modern highway toll systems. This paper presents an embedded technology-based License Plate Recognition System. The system uses a multi-processor technology, with three DSP and two FPGA included. Through a multi-block high-speed processor application, the accuracy of recognition system will be at a very high level. The greatest advantage of DSP and FPGA System lies in its structure flexibility, high universality and fitting in the modular design. Therefore, it can achieve highly efficient three sets of algorithm and real-time control. At the same time, its development process can be carried out in parallel.

As a novel information acquiring and processing technology, compared to other traditional sensor networks, wireless multimedia sensor networks pay more attention to the information-intensive data (e.g. audio, video, image). Potential applications of wireless multimedia sensor networks span a wide spectrum from military to industrial, from commercial to environmental monitoring. This paper discusses QoS routing problem model with multiple QoS constraints, which may deal with the delay, delay jitter and bandwidth, and presents QoS routing algorithm with multiple constraints based on genetic algorithm, and gives the algorithm idiographic flow. Simulating results show that higher routing successful ratio and faster convergence is achieved in the algorithm.

Along with the rapid development of biometric identification techniques, the fingerprint identification is becoming a significant subject. Automated fingerprint identification is a method to identify a person based on his fingerprint physiological characteristics. This paper presents an automatic fingerprint capture and preprocessing system with a fixed point DSP, TMS320VC5510A and a fingerprint sensor, FPC1011C. This system supplies two types of power: 1) wall adapter power, which can automatically switch from the wall adapter power to the battery in case of power-fail or brownout conditions, and 2) battery . With the interactive module , keyboard and the LCD in this system, the algorithm can run smoothly to realize the fingerprint processing.

The localization technology based on beacon-free node can significantly reduce the network cost when it takes into account the configuration constraints and cost factors of wireless sensor network as well as satisfies the lower hardware requirements. This inspires the wireless sensor network node localization studies. With an analysis of the beacon nodes and beacon-free node localization algorithm, this paper puts forward the localization algorithm of wireless sensor network beacon-free node probability-based. Simulation results show that the algorithm, at no additional hardware, will reduce the computation overhead and improve the localization accuracy and that has good adaptability.

In all intelligent systems, knowledge is the cornerstone of the construction of system. This paper discusses a number of ways of the knowledge expression, such as rule, processing, semantic network, nonspecification logic ,as well as the organization, management and acquisition of knowledge. With the electrical apparatus products as an example, the paper proposes the approach of designing object knowledge, including forward and reverse mixed reasoning mechanism and rule based reasoning(RBR) method under the fuzzy mechanism. It also presents the establishment of an intelligent model of knowledge processing and evaluation method for projects of product conceptual design.