Artificial Intelligence and Computational Intelligence

According to the shortcoming that the Extended Kalman filter (EKF) method can only estimate the speed and rotor position of induction motors in time domain when it is used to diagnose the fault existed in induction motor. A new multi-scale default diagnosing method is developed by combining EKF and wavelet transform. By monitoring the voltages and currents of the stator, it is possible to estimate the speed and position on-line. The new filter combines the merit of EKF and wavelet, and it not only possesses the multiscale analysis capability both in time domain and frequency domain, but also has better estimation accuracy than traditional EKF. Computer simulation shows the effect of the new algorithm.

In recent years, computational intelligence methods are widely used to solve problems in engineering structural field by more and more researchers. In this paper, a method based on genetic algorithm (GA) for identifying the damage in a roof truss under static loads has been developed. At first, the forward analysis based on the finite element model method clearly demonstrates that the damage of elements on a roof truss can result in a change of static axial strain. Then GA has been used to identify the location and the degree of structural damage. In this paper, damage in the structure is modeled as a reduction in the cross-sectional of the damaged element. The identification problem is formulated as a constrained optimization problem, which is solved using GA. Unlike the traditional mathematical methods, which guide the direction of hill climbing by the derivatives of objective functions, GA searches the problem domain by the objective functions itself at multiple points. The objective function is defined as the difference between the measured static strains and the analytically predicted strains obtained from a finite element model. By minimizing the objective function, the damage location and damage severity can be successfully identified. The static-based method uses only strain measurements at a few degrees of freedom as input to the identification procedure and no additional excitation is required. These features make the method ideally suited for long-term monitoring of large civil structures. The method proposed in this paper is demonstrated using a plane roof truss model, and the results fit well with the actual value.

To solve the imprecise description of efficiency and improve the accuracy of the key hydraulic components models in complex operating conditions, the traditional hydraulic pump and motor model is discussed and improved. With the non-linear improvement and the parameter optimization algorithms, model parameters can be determined based on the experimental efficiency data of samples. Take a motor product sample for example, the efficiency distribution of the improved model is much closer to the experimental results than that of the traditional model. The mean value and the variance value of percentage error for the improved model are much smaller, and the error analysis proves that the improved model is much more suitable for the modeling in complex operating conditions.

In this paper the objective was to study diurnal and seasonal changes in Yulan Magmolia tree by SWR principle. Laboratory and field tests were performed. Laboratory calibration test was performed on two sapwood samples, which shows that the relation between the volumetric water content and the output voltage of SWR sensor was monotonic as the coefficients R

2

reached above 0.90.were periodically weighed on a balance. In field test, the diurnal and seasonal changes were monitored in natural Yulan Magmolia tree by SWR sensor for nearly 2 years. It indicated that seasonal variation of stem water content measured was 20% ~ 25%. A maximum in stem water content occurred in summer, then decreased in autumn, and reaching a minimum in winter, recovery is almost complete till next spring. The short-term (diurnal) variations were found 2% ~ 5%. The daily changes show that a decline in stem water content happened, and then recovered each day. At the same time, leaf water content and sap flow velocity were measured, which shows that diurnal leaf water potential has the similar curve with diurnal stem water content while sap flow velocity was reverse to stem water content.

Reliability analysis on the aircraft structures is an integrative study about the structural components and the endured force load on aircraft structures. Especially, the wing reliability is an important index of the aircraft reliability. The reliability of a wing structure model under gust load is analyzed by computer simulation in this paper using the Probability Finite Element Method (PFEM). The gust load, elastic modulus and yield strength are taken as input variables to simulate the reliability problem of the wing structures using the Monte Carlo method, and the influences of the input parameters on aircraft wing strength are obtained. This may provide a viable analysis method for the design and manufacture of an aircraft.

In spacecraft telemetry data prediction field, unknown residual distribution and great volatility of predicted value have hampered traditional prediction interval methods to follow forecast trend and give high-precision intervals. Hence, modified Block Bootstrap prediction interval Method is proposed in this paper. Contrast to traditional method, this method can enhance accuracy of non-stationary time series data prediction interval for its data sampling frequency can be adjusted by data character. In the end, an example is given to show the validity and practicality of this method.

To study the issues of network coverage and network of life in the three-dimensional geographical environment, the cross deployment strategies and random deployment strategies were adopted, and random-sleep dissemination algorithm based on the dynamic neighbors’ nodes information was used to study the regional coverage, node coverage and the network life under these two strategies. The experiment results show that, after using random-sleep dissemination algorithm based on the dynamic neighbors nodes information, the effect of the regional coverage throughout the region changes little, the network life extends about 1.1 times when there are enough nodes.

In this study the underlying dimensions of e-commerce benefits as perceived by businesses are investigated. Statistical analysis of empirical data from a questionnaire survey of managers of Chinese businesses resulted in three meaningful e-commerce benefits dimensions being uncovered. They are the Market, Efficiency, and Quality dimensions, which represent what companies perceive as business aspects where they can derive benefits by adopting e-commerce. In addition, the limitations of this study and directions for future research are discussed.

An extension of an optimal velocity model with the relative velocity is analyzed. From the nonlinear analysis, the propagating kink solution for traffic jams is obtained. The fundamental diagram and the relation between the headway and the delay time are examined by numerical simulation. We find that the result from the nonlinear analysis is in good agreement with that obtained from the numerical simulation.

In this study, a training computer simulation system for acupuncture training was proposed for the purpose of a quantitative characterization of the traditional oriental technique. A system using a force measuring device was constructed to record and analyze the basic data of insertion force. The index of insertion force was decided, and fundamental data for the development of such a computer simulation system of acupuncture training was obtained.

There often exist redundant channels in EEG signal collection which deteriorate the classification accuracy. In this paper, a classification method which can deal with redundant channels, as well as redundant CSP features, is presented for motor imagery task. Our method utilizes CSP filter and margin maximization with linear programming to update a compound weight that enables iterative channel elimination and the update of the following linear classification. Theoretical analysis and experimental results show the effectiveness of our method to solve redundancy of channels and CSP features simultaneously when classifying motor imagery EEG data.

Automatic Electroencephalography (EEG) interpretation system had been developed as an important tool for the inspection of brain diseases. In this study, an automatic reference selection technique was developed to obtain the appropriate derivation for EEG components interpretation. The cross spectrum of bipolar EEG was adopted to detect the phase reversal among the EEG channels covering the scalp of head. Appropriate reference was selected automatically based on the detected phase reversal. Finally, a referential derivation was constructed. The distribution of EEG components was analyzed based on the constructed referential derivation to evaluate the effectiveness of selected reference for quantitative EEG component interpretation.

A novel mixed numerical integral algorithm is proposed for the deformation simulation of soft tissues in virtual surgery system. First, a Mass-Spring System is built as the soft tissue’s kinetic model. And then, a mixed numerical integral algorithm is derived to solve the model’s deformation differential equations. At the end of the paper, results are presented as examples, which validate the effectiveness of our algorithm in improving simulation performance by reducing complexity and enhancing accuracy.

In this paper, we apply the artificial bee colony (ABC) and its improving format to solve multiple sequence alignment (MSA) problem. The improved method named ABC_SA, which is presented to prevent algorithm from sliding into local optimum through introducing Metropolis acceptance criteria into ABC’s searching process. The results of simulation experiment demonstrate that ABC_SA algorithm is able to settle multiple sequence alignment effectively by increasing the food source’s diversity and is able to converge at global optimal alignment.

With the fast development of Radio Frequency Identification (RFID) technology, RFID network has been applied in different aspects of logistic management. How to effectively deploy the readers becomes a crucial problem in RFID network planning. The planning is related to a complicated optimization problem and interference elimination between readers. To find a good solution in the optimization problem effectively, we introduced Differential Evolution algorithm. To minimize the interference between the readers, we applied TDMA on the network and proposed two methods to group the readers. The first method is a modified version of Differential Evolution algorithm. Since part of the problem domain is binary while the searching space of the Differential Evolution algorithm is in a real domain, we modified the mutation rule of the Differential Evolution algorithm so that it can support binary parameters. The other way is to transform the problem into a graph and apply a maximum cut heuristic on it. The experimental result shows that both methods are effective.

Diagnosis of reaction wheel faults is very significant to ensure long-term stable satellite attitude control system operation. Support vector machine (SVM) is a new machine learning method based on statistical learning theory, which can solve the classification problem of small sampling, non-linearity and high dimensionality. However, it is difficult to select suitable parameters of SVM. Particle Swarm Optimization (PSO) is a new optimization method, which is motivated by social behavior of bird flocking. The optimization method not only has strong global search capability, but is also very simple to apply. However, PSO algorithms are still not mature enough for handling some of the more complicated problems as the one posed by SVM. Therefore an improved PSO algorithm is proposed and applied in parameter optimization of support vector machine as IPSO-SVM. The characteristics of satellite dynamic control process include three typical reaction wheel failures. Here an IPSO-SVM is used in fault diagnosis and compared with neural network-based diagnostic methods. Simulation results show that the improved PSO can effectively avoid the premature phenomenon; it can also optimize the SVM parameters, and achieve higher diagnostic accuracy than artificial neural network-based diagnostic methods.

Using the global optimization properties of Particle Swarm Optimization(PSO) to carry out parameter identification of support vector machine(SVM). Before the particle swarm search for parameters, exponential transform the parameters first to make intervals [0, 1] and [1, infinity] have the same search probability. Fitness function of PSO as generalization ability of support vector machine model to be the standard, at the same time discussed the minimum error of testing samples and leave-one-out method to the SVM learning method promotion ability. Finally taking the data of monthly runoff of Yichang station in Yangtze River as an example, respectively using the ARMA model, seasonal ARIMA model, BP neural network model and the SVM model that have built to simulate forecasting, the result shows the validity of the model.

Based on Latin Hypercube Sampling method, the crossover operation in GA is redesigned; combined with immune mechanism, chromosome concentration is defined and clonal selection strategy is designed, thus an Immune Genetic Algorithm is given based on Latin Hypercube Sampling for solving the Maximum Clique Problem in this paper. The examples shows the new algorithm in solution quality, convergence speed, and other indicators is better than the classical genetic algorithm and good point set genetic algorithm. On the other hand, the new algorithm is not inferior to such classical algorithms as dynamic local search, and it got better solutions to some examples.

The sales manager selection is very important for human resource planning of enterprises because it directly affects their service quality. That is to say, an appropriate sales manager is able to improve communication and provide necessary services for customers. Thus, sale ability of an enterprise will be increased by appropriate sales managers. To select appropriate sales managers easily and quickly, we utilize a fuzzy multi-criteria decision-making(FMCDM) method called fuzzy TOPSIS to solve the selection problem in enterprises. Fuzzy TOPSIS is an extension of TOPSIS under fuzzy environment, and TOPSIS is one of famous multi-criteria decision-making(MCDM) methods under certain environment. By fuzzy TOPSIS, an optimal sales manager is easily found from lots of candidates. Besides, selection criteria of the FMCDM method will be also evaluation standards of sales managers’ ability after worked in the enterprise.

The impact of the random sequence on Genetic Algorithms (GAs) is rarely discussed in the community so far. The requirements of GAs for Pseudo Random Number Generators (PRNGs) are analyzed, and a series of numerical experiments of Genetic Algorithm and Direct Search Toolbox computing three different kinds of typical test functions are conducted.An estimate of solution accuracy for each test function is included when six standard PRNGs on MATLAB are applied respectively. A ranking is attempted based on the estimated solution absolute/relative error. It concludes that the effect of PRNGs on GAs varies with the test function; that generally speaking, modern PRNGs outperform traditional ones, and that the seed also has a deep impact on GAs. The research results will be beneficial to stipulate proper principle of PRNGs selection criteria for GAs.

This paper aims to rank LR-fuzzy numbers (LR-fns) by the pairwise comparison based method. Different from the existing methods, our method uses the information contained in each LR-fn to get a consistent total order. In detail, since an LR-fn may not be absolutely larger or smaller than another, this paper proposes the concept of dominant degree to quantify how much one LR-fn is larger and smaller than another. From the dominant degrees, we construct a pairwise comparison matrix based on which a consistent ranking is got. Meanwhile, the ranking result is transitive and consistent, and agrees with our intuition. Examples and comparison with existing methods show the good performance of our method.

According to the principle of theory combining with practice, preliminary indicators of road safety evaluation are selected by integrating the characteristics of road traffic safety in China. On the basis of preliminary indicators, evaluation index system of traffic safety is constructed from three respects: accident rate, accident severity and the improved level of safety management. Considering the fuzziness, dependence and interaction of assessing indices, we applied the fuzzy method to assess traffic safety, and the evaluation and grading criteria for different indexes were presented. Finally, the practical value of the evaluation system and the evaluation model proposed in this study were proved by case study. Furthermore, the results of the study may provide reference for the traffic management departments.

Construction site layout is crucial to any project and has a significant impact on the economy, quality, schedule, and other aspects of the project. Especially for infrastructure construction project, the layout planning of site temporary batch plant is one of the key factors for assessing on complexity scales and project performances. The poorer the site batch plant layout planning, the more increasing the project complexity (risks) and the greater the construction cost raised. Generally, the site batch plant layout problem has been solved through the experiences of site management team, using more or less sophisticated numerical model, but mostly the final decision made by the leader who responsible for the site management taking much more consideration of the design conditions and the site spatial constrained. Therefore, the site batch plant layout has always been considered in complying with maximization principles. That is, individual contractor is often required to build his own batch plants for servicing his own project only according to its contract clauses. This construction management strategy has caused to a large waste of resources and much low productivity of the operated plants. The purpose of this paper applies genetic algorithms and computing techniques to develop an optimal model in order to search an optimal solution to the site batch plant layout in the planning stage, and combine with the practical contracting strategies to design feasible construction management plan for enhancement of site management and improvement of concrete quality, as well as minimizing the total construction costs. The GA’s model was developed and applied to specific project located in Taiwan. The usefulness of the model was proven through by the practical operation of the project.

An algorithm based on damped vibration for multi-objective optimization problems is proposed in this paper. This algorithm makes use of the concept of damped vibration to do local search to find optimal solutions. The concept of Pareto Dominance is used to determine whether a solution is optimal. Meanwhile, the use of many random vibrators and the randomness of the initial maximum displacement ensure that the solutions are global. Simulation results show that the damping search algorithm is efficient in finding more solutions and also have good convergence and solution diversity.

Genetic algorithm is known as one of the ways of resolving complicated problems and optimization issues. This algorithm works based on a search space and in this space it’d seeking for the optimum answer. In this algorithm, there exist agents and gorges which expand the search space with no logical reason. We can find the measures which take us away from the optimal answer by observing the trend of changes, and it can apply the changes in a way that increases the speed of reaching the answers. It’s obvious these changes must be as much as they don’t add time complexity or memory load to the system.

In this paper, we represent a mechanism as a pruning operator in order to reach the answer more quickly and make it work optimal by omitting the inappropriate measures and purposeful decrease of search space.

Many issues in engineering computation and practical application that ultimately boil down to a matrix computation. And different applications will lead to some of the special sparse structure of the matrix computation. A modified chasing method has been proposed to solve the sevendiagonal linear equations in this paper firstly. By using this method, the condition that each principal minor sequence of coefficient matrix must nonzero is unnecessary. At the same time, we present a new computational algorithm for solving periodic sevendiagonal linear systems. An example is given in order to illustrate the algorithm.

For continuous processes, the global LSSVM always gives good prediction for testing data located in the neighborhood of dense training data but incompetent for these in the sparse part. To solve the problem, the paper proposed a local weighted LSSVM method in the online modeling of continuous processes. At each period, only the samples similar to the current input are added into the training set and the obtained model is just for predicting the current output. To distinguish the importance of the training data, weight is defined to each data according to the Euclidean distances between the training data and testing data. The presented algorithm is applied in pH neutralization process and the result shows the excellent performance of the presented algorithm in precision and predicting time.

This paper presents a Cellular Automata (CA) based crowd behavior model which mimics movements of humans in an indoor environment. Because of the increasing population in modern cities, the understanding of crowd behavior in the urban environment has become a crucial issue in emergency management. In the proposed crowd behavior model, pedestrians are confined to move in a cellular space where their movements are determined by their own status and the surrounding environment characteristics. A pedestrian’s behavior is constructed from several attributes: including the “walking toward goal” behavior, “collision and congestion avoidance” behavior, “grouping” behavior and path smoothness. Simulations have been carried out with a crowd consisting of thirty pedestrians in an indoor environment to validate the model.

In this paper, a new approach for protecting the ownership of relational database is presented. Such approach is applied for protecting both textual and numerical data. This is done by adding only one hidden record with a secret function. For each attribute, the value of this function depends on the data stored in all other records. Therefore, this technique is more powerful against any attacks or modifications such as deleting or updating cell values. Furthermore, the problems associated with the work in literature are solved. For example, there is no need for additional storage area as required when adding additional columns especially with large databases. In addition, in case of protecting data by adding columns, we need to add a number of columns equal to the number of data types to be protected. Here, only one record is sufficient to protect all types of data. Moreover, there is a possibility to use a different function for each field results in more robustness. Finally, the proposed technique does not have any other requirements or restrictions on either database design or database administration.

Moving a brain-computer interface from a laboratory demonstration to real-life applications poses severe challenges to the BCI community. Recently, with advances in the biomedical sciences and electronic technologies, the development of a mobile and online BCI has obtained increasing attention in the past decade. A mobile and wireless BCI based on customized Electroencephalogram recording and signal processing modules has the advantage of ultimate portability, wearability and usability. This study integrates a mobile and wireless EEG system and a signal-process platform based on a Bluetooth-enabled cell-phone into a truly wearable BCI. Its implications for BCI were demonstrated through a case study in which the cell-phone was programmed to assess steady-state, visual-evoked potentials in response to flickering visual stimuli to make a phone call directly. The results of this study on ten normal healthy volunteers suggested that the proposed BCI system enabled EEG monitoring and on-line processing of unconstrained subjects in real-world environments.

This paper studies a script language DCISL for the dynamic control integration of the MAS (Multi-Agent System), which is used to describe the integration rules based on service flow. With the analysis of MAS-related language elements and combination of formalized description method and structure tool, this paper provides the definition of DCISL and realization of DCISL interpreter. DCISL uses the way of unifying the logistic definition and separating the physical realization to describe the MAS integration rules. Through the script interpreter, the integration rules between service agents which are described in DCISL are mapped into behavior logic between service agents, and by using the interpreting mechanism of the script interpreter, the dynamic script switch is realized. Through the interpreter within the service agents, the integration rules within the agents which are described in DCISL are mapped into the behavior logic between function agents, and according to the contract net protocol, the service agents interact with CMB (Common Message Blackboard) to realize the dynamic bid inviting. Ultimately, the experiment shows that the agents can run independently according to the integration logic based on the integration rules which are described in DCISL and eventually realizes the cooperation between agents and the dynamic integration control of agents system by using of the script switch and bid inviting mechanism.

Multi-view architecture products describe the requirement connotation of complex information system explicitly and reflect the essence of problem domain. Agent organization gives a good idea in solving problem domain in course of its integrated and intelligent. This paper takes the DOD AF operational view products for example to combine with their merits. After the formalized expression of operational view products relating to organization abstract model, some mapping rules to mapping operational view products to organization abstract model were suggested. Then using the social network analysis method, a way of mapping organization abstract model to multi-agent software architecture style according to Pearson correlation coefficients was set up. In the end, an example was set up to validate the feasibility of this approach. This work can integrate the system requirements and multi-agent software architecture. It shortened the gap between system’s top-design and its implementation, at the same time, it improved the system adaptability.

It has became a new active topic in grid security research fields that authentication and keys agreement between heterogeneous and dynamic grid nodes which belong to different trust domains. In this paper, an ID-based authenticated multi-group keys agreement scheme is proposed from bilinear pairings. Proposed scheme provides mutual authentication method for users that belong to different trust domains, and it employs shared password evolvement authentication mechanism that generates one-time passwords for every session. Moreover, our scheme overcomes limitations existed in some existing pairings-based authentication protocol, in which, special Hash functions are necessary. Proposed scheme can be used to generate session keys for large-scale distributed heterogeneous and dynamic grid resource.

For the global path planning of mobile robot under the dynamic uncertain environment, a path planning method combined time rolling window strategy and artificial bee colony (ABC) algorithm was proposed. To meet the real time requirement, the global path was replaced by local paths within a series of rolling windows. Due to the ability of global optimization, and rapid convergence of artificial bee colony algorithm, it was applied to plan the local path. According to the special environment, a suitable fitness function was designed to avoid dynamic obstacles in artificial bee colony algorithm. The simulation results of proposed method demonstrated that it has great efficiency and accuracy, and it is suitable for solving this kind of problem.

To optimize urban arterial traffic control, this paper analyzed coordination mechanism of all individual junctions along the road. We set up a traffic control system for urban area network based upon multi-agent technology. Each individual junction and the coordination were considered as agents. Each of them was embodiment of fuzzy neural network. We utilized particle swarm optimization arithmetic to optimize these FNNs. The agent directly talked to each other with FIPA ACL standard language. Compared to the traditional timing control mode, at a junction with moderate traffic volume, the average delay expressed in queue length can be reduced from 120.9(veh./h) to 25.4 (veh./h). Congestion thus significantly relieved.

Rainfall forecasting is very important research topic in disaster prevention and reduction. In this study, a semiparametric regression ensemble (SRE) model is proposed for rainfall forecasting based on radial basis function (RBF) neural network. In the process of ensemble modeling, original data set are partitioned into some different training subsets via Bagging technology. Then a great number of single RBF neural network models generate diverse individual neural network ensemble by training subsets. Thirdly, the partial least square regression (PLS) is used to choose the appropriate ensemble members. Finally, SRE is used for neural network ensemble for prediction purpose. Empirical results obtained reveal that the prediction using the SRE model is generally better than those obtained using the other models presented in this study in terms of the same evaluation measurements. Our findings reveal that the SRE model proposed here can be used as a promising alternative forecasting tool for rainfall to achieve greater forecasting accuracy and improve prediction quality further.

This paper proposes a simple and effective modified particle swarm optimizor with a novel operator. The aim is to prevent premature convergence and improve the quality of solutions. The standard PSO is shown to have no ability to perform a fine grain search to improve the quality of solutions as the number of iterations is increased, although it may find the near optimal solutions much faster than other evolutionary algorithms. The modified PSO algorithm presented in this paper is able to find near optimal solutions as fast as the standard PSO and improve their quality in the later iterations. Compared with the standard PSO, benchmark tests are implemented and the result shows that our modified algorithm successfully prevents premature convergence and provides better solutions.

In order to keep balance of premature convergence and diversity maintenance, an AntiCentroid-oriented particle updating strategy and an improved Particle Swarm Algorithm (ACoPSA) are presented in this paper. The swarm centroid reflects the search focus of the PSA algorithm and its distance to the global best particle (gbest) indicates the behavior difference between the population search and the gbest. Therefore the directional vector from the swarm centroid to the gbest implies an effective direction that particles should follow. This direction is utilized to update the particle velocity and to guide swarm search. Experimental comparisons among ACoPSA, standard PSA and a recent perturbed PSA are made to validate the efficacy of the strategy. The experiments confirm us that the swarm centroid-guided particle updating strategy is encouraging and promising for stochastic heuristic algorithms.

Multidisciplinary Design Optimization (MDO) is an effective and prospective solution to complex engineering systems. In MDO methodology, MDO algorithm is the most important research area. Four decomposition algorithms have been proposed for MDO. They are Concurrent subspace optimization (CSSO), Collaborative optimization (CO), Bi-level integrated system synthesis (BLISS) and Analytical target cascading (ATC). On the basis of specific requirements for comparison, a mathematical example is chose and the performances of MDO decomposition algorithms are evaluated and compared, which take into consideration optimization efficiency and formulation structure characteristics.

Image thresholding is an important technique for image processing and pattern recognition. The maximum entropy thresholding (MET) has been widely applied. A new multilevel MET algorithm based on the technology of the artificial bee colony (ABC) algorithm is proposed in this paper called the maximum entropy based artificial bee colony thresholding (MEABCT) method. Three different methods, such as the methods of particle swarm optimization, HCOCLPSO and honey bee mating optimization are also implemented for comparison with the results of the proposed method. The experimental results manifest that the proposed MEABCT algorithm can search for multiple thresholds which are very close to the optimal ones examined by the exhaustive search method. Meanwhile, the results using the MEABCT algorithm is the best and its computation time is relatively low compared with other four methods.

While fuzzy logic controllers (FLCs) are developed to exploit human expert knowledge in designing control systems, the actual establishment of fuzzy rules and tuning of fuzzy membership functions are usually a time consuming exercise. In this paper a technique, based on the particle swarm optimisation (PSO), is employed to automatically tune the fuzzy rules of a Mamdani-type of fuzzy controller. The effectiveness of the designed controller is demonstrated by the control performance of such an FLC to a nonlinear water tank system with process time delay. The results are compared favourably to a PSO tuned PID controller.

We develop an efficient differential evolution (DE) with neural networks-based approximating technique for computationally expensive problems, called DE-ANN hereinafter. We employ multilayer feedforward ANN to approximate the original problems for reducing the numbers of costly problems in DE. We also implement a fast training algorithm whose data samples use the population of DE. In the evolution process of DE, we combine the individual-based and generation-based methods for approximate model control. We compared the proposed algorithm with the conventional DE on three benchmark test functions. The experimental results showed that DE-ANN had capacity to be employed to deal with the computationally demanding real-world problems.

The software quality model and software quality measurement model are the basis of evaluating the quality of the Foundational Software Platform (FSP), but it is quite difficult or even impossible to collect the whole metric data required in the process of the software quality measurement, which is the problem of the FSP quality evaluating. Bayesian networks are the suitable model of resolving the problem including uncertainty and complexity. By analyzing the problem domain of foundational software platform quality evaluation and comparing it with the characteristic domain of Bayesian networks, this paper proposed a method of evaluating the quality of the FSP by Bayesian network. The method includes three parts: node choosing, Bayesian network learning and Bayesian network inference. The results of the experiments indicate a Bayesian network for every quality characteristic should be built in practical quality evaluation of the FSP by the proposed method.

The method on uncertain multiple attribute group decision making (MAGDM) based on aggregating intuitionistic fuzzy information is investigated.Firstly,some operational laws,score function and variation function of triangle fuzzy number intuitionistic fuzzy sets(TFNIFSs) are proposed. Then,triangle fuzzy number intuitionistic fuzzy weighted geometric (TFNIFWG) operator and triangle fuzzy number intuitionistic fuzzy ordered weighted geometric (TFNIFOWG) operator are studied. Further, a TFNIFWG and TFNIFOWG operators-based approach is developed to solve the MAGDM problems in which the attribute values take the form of TFNIFSs and the expert weights are completely unknown.Finally, some illustrative example is given to verify the developed approach and to demonstrate its practicality and effectiveness.

This paper presents new results on the statistical analysis of wireless fading channels. Exact closed-form expressions are derived for average Doppler Shift and average Doppler Spread, Probability density function of received signal’s amplitude and phase, level crossing rate and average fading interval, and probability distribution of fading interval. The utility of the new theoretical formulas, validated by Monte Carlo simulations, are briefly discussed as well.

The attributes in rough set must be discretized, but the general theory on discretization did not think about the decision attribute adequately during discretization of data, as a result, it leads to several redundant rules and lower calculation efficiency. The discretization method of continuous attributes based on decision attributes which is discussed in this paper gives more attention to both significance of attributes and the decision attributes. The continuous attributes are discretized in sequence according to their significance. The result shows less breakpoints and higher recognition accuracy. The experiment on database Iris for UCI robot learning validates the feasibility of our method. Comparing the result with documents [6] and [11], the method given in this paper shows higher recognition accuracy and much less breakpoints.

In theory, a gold futures possesses function of price discovery. However, futures including information must be disclosed by some effective way. This paper proposes a forecasting model which combines wavelet frame with Support vector regression (SVR). Wavelet frame is first used to decompose the series of gold futures price into sub-series with different scales, the SVR then uses the sub-series to build the forecasting model. Empirical research shows that the gold futures has the function of price discovery, and the two steps model is a good tool for making the price information clear and forecasting spot price. further research can try different basis function or other methods of disclosing information.