Advances in Swarm Intelligence

This paper considers the growth rate dynamics of a predator-prey system as a discrete event dynamical system. Timed Petri nets are a graphical and mathematical modeling tool applicable to discrete event systems in order to represent its states evolution. Lyapunov stability theory provides the required tools needed to aboard the stability problem for the predator-prey system treated as a discrete event system modeled with timed petri nets. By proving boundedness one confirms a dominant oscillating behavior of both populations dynamics performance. However, the oscillating frequency results to be unknown. This inconvenience is overcome by considering a specific recurrence equation, in the max-plus algebra.

Particle swarm optimization (PSO) has been shown to perform well on many optimization problems. However, the PSO algorithm often can not find the global optimum, even for unimodal functions. It is necessary to study the local search ability of PSO. The interval compression method and the probabilistic characteristic of the searching interval of particles are used to analyze the local search ability of PSO in this paper. The conclusion can be obtained that the local search ability of a particle is poor when the component of the global best position lies in between the component of the individual best position and the component of the current position of the particle. In order to improve the local search ability of PSO, a new learning strategy is presented to enhance the probability of exploitation of the global best position. The experimental results show that the modified PSO with the new learning strategy can improve solution accuracy.

For improving the search performance of a canonical particle swarm optimizer (CPSO), we propose a newly canonical particle swarm optimizer with diversive curiosity (CPSO/DC). A crucial idea here is to introduce diversive curiosity into the CPSO to comprehensively manage the trade-off between exploitation and exploration for alleviating stagnation. To demonstrate the effectiveness of the proposed method, computer experiments on a suite of five-dimensional benchmark problems are carried out. We investigate the characteristics of the CPSO/DC, and compare the search performance with other methods. The obtained results indicate that the search performance of the CPSO/DC is superior to that by EPSO, ECPSO and RGA/E, but is inferior to that by PSO/DC for the

Griewank

and

Rastrigin

problems.

Bee-Swarm genetic algorithm based on reproducing of swarm is a novel improved genetic algorithm. Comparing to GA, there are two populations, one for global search, and another for local search. Only best one can crossover. The genetic operators include order crossover operator, adaptive mutation operator and restrain operator. The simulated annealing is also introduced to help local optimization. The method sufficiently takes the advantage of genetic algorithm such as group search and global convergence, and quick parallel search can efficiently overcome the problems of local optimization. Theoretically, the capability of finding the global optimum is proved, and a necessary and sufficient condition is obtained namely. The convergence and effective of BSGA is proved by Markov chain and genetic mechanism. Finally, several testing experiments show that the Bee-Swarm genetic algorithm is good.

Artificial Searching Swarm Algorithm (ASSA) is an intelligent optimization algorithm, and its performance has been analyzed and compared with some famous algorithms. For farther understanding the running principle of ASSA, this work discusses the functions of three behavior rules which decide the moves of searching swarm. Some typical functions are selected to do the simulation tests. The function simulation tests showed that the three behavior rules are indispensability and endow the ASSA with powerful global optimization ability together.

We describe the general foraging task, breaking it into two different subtasks: map-making and collection. Map-making is a task in which a map is constructed which contains the location(s) of an item or of items in the search area. Collection is the task in which an item is picked up and carried back to a central known location. We theoretically examine these tasks, generating minimal conditions for each one to be accomplished. We then build a swarm made up of two castes to accomplish this, theoretically motivating the design of the swarm. Finally, we demonstrate that the swarm is optimal in the class of swarms utilizing line-of-sight communication, and give performance measures for open and closed search spaces.

This study proposes an adaptive staged particle swarm optimization (ASPSO) algorithm based on analyses of particles’ search capabilities. First, the search processes of the standard PSO (SPSO) and the linear decreasing inertia weight PSO (LDWPSO) are analyzed based on our previous definition of exploitation. Second, three stages of the search process in PSO are defined. Each stage has its own search preference, which is represented by the exploitation capability of swarm. Third, the mapping between inertia weight, learning factor (

w

-

c

) and the exploitation capability is given. At last, the ASPSO is proposed. By setting different values of

w

-

c

in three stages, one can make swarm search the space with particular strategy in each stage, and the particles can be directed to find the solution more effectively. The experimental results show that the proposed ASPSO has better performance than SPSO and LDWPSO on most of test functions.

The speed equation of particle swarm optimization is improved by using a convex combination of the current best position of a particle and the current best position which the whole particle swarm as well as the current position of the particle, so as to enhance global search capability of basic particle swarm optimization. Thus a new particle swarm optimization algorithm is proposed. Numerical experiments show that its computing time is short and its global search capability is powerful as well as its computing accuracy is high in compared with the basic PSO.

Based on the analysis of biological symbiotic relationship, the mechanism of facultative parasitic behaviour is embedded into the particle swarm optimization (PSO) to construct a two-population PSO model called PSOPB, composed of the host and the parasites population. In this model, the two populations exchange particles according to the fitness sorted in a certain number of iterations. In order to embody the law of "survival of the fittest" in biological evolution, the poor fitness particles in the host population are eliminated, replaced by the re-initialization of the particles in order to maintain constant population size. The results of experiments of a set of 6 benchmark functions show that presented algorithm model has faster convergence rate and higher search accuracy compared with CPSO, PSOPC and PSO-LIW.

It is not trivial to tune the swarm behavior just by parameter setting because of the randomness, complexity and dynamic involved in particle swarm optimizer (PSO). Hundreds of variants in the literature of last decade, brought various mechanism or ideas, sometimes also from outside of the traditional metaheuristics field, to tune the swarm behavior. While, in the same time, additional parameters have to be afforded. This paper proposes a new mechanism, named KNOB, to directly tune the swarm behavior through parameter setting of PSO. KNOB is defined as the first principal component of the statistical probability sequence of exploration and exploitation allocation along the search process. The using of the KNOB to tune PSO by parameter setting is realized through a statistical mapping, between the parameter set and the KNOB, learned by a radial basis function neural network (RBFNN) simulation model. In this way, KNOB provides an easy way to tune PSO directly by its parameter setting. A simple application of KNOB to promote is presented to verify the mechanism of KNOB.

This paper proposes a novel population-based evolution algorithm named grouping-shuffling particle swarm optimization (GSPSO) by hybridizing particle swarm optimization (PSO) and shuffled frog leaping algorithm (SFLA) for continuous optimization problems. In the proposed algorithm, each particle automatically and periodically executes grouping and shuffling operations in its flight learning evolutionary process. By testing on 4 benchmark functions, the numerical results demonstrate that, the optimization performance of the proposed GSPSO is much better than PSO and SFLA. The GSPSO can both avoid the PSO’s shortage that easy to get rid of the local optimal solution and has faster convergence speed and higher convergence precision than the PSO and SFLA.

The paper presents a novel particle swarm optimizer (PSO), called gender-hierarchy particle swarm optimizer based on punishment (GH-PSO). In the proposed algorithm, the social part and recognition part of PSO both are modified in order to accelerate the convergence and improve the accuracy of the optimal solution. Especially, a novel recognition approach, called general recognition, is presented to furthermore improve the performance of PSO. Experimental results show that the proposed algorithm shows better behaviors as compared to the standard PSO, tribes-based PSO and GH-PSO with tribes.

This paper deals with the problem of unconstrained optimization. An improved probability particle swarm optimization algorithm is proposed. Firstly, two normal distributions are used to describe the distributions of particle positions, respectively. One is the normal distribution with the global best position as mean value and the difference between the current fitness and the global best fitness as standard deviation while another is the distribution with the previous best position as mean value and the difference between the current fitness and the previous best fitness as standard deviation. Secondly, a disturbance on the mean values is introduced into the proposed algorithm. Thirdly, the final position of particles is determined by employing a linear weighted method to cope with the sampled information from the two normal distributions. Finally, benchmark functions are used to illustrate the effectiveness of the proposed algorithm.

Niching is an important technique for mutlimodal optimization. This paper proposed an improved niching technique based on particle swarm optimizer to locate multiple optima. In the proposed algorithm, the algorithm inspired from natural ecosystem form niches automatically without any prespecified problem dependent parameters. Experiment results demonstrated that the proposed niching method is superior to the classic niching methods which are with or without niching parameters.

A major challenge in task scheduling is the availability of resources. In a heterogeneous environment, where processors operate at different speeds and are not continuously available for computation, achieving a better make-span is a key issue. The existing algorithm SSAC has proved to be a good trade-off between availability and responsiveness while maintaining a good performance in the average response time of multiclass tasks. But the makespan may be influenced due to load imbalance. In this paper we proposed approach try to further optimize this scheduling strategy by using quantum-behaved particle swarm optimization. And compared with SSAC and MINMIN in the simulation experiment; results indicate that our proposed technique is a better solution for reducing the makespan considerably.

This paper investigates the two-machine flow shop group scheduling problem with the transportation times and sequence-dependent setup times considerations. The objective is to minimize the total completion time. In this paper, a novel encoding scheme of PSO for flow shop group scheduling is proposed to effectively solve various instances with group numbers up to 15. Note that the proposed encoding scheme simultaneously determines the sequence of jobs in each group and the sequence of groups. Three different lower bounds are developed to evaluate the performance of the proposed PSO algorithm. Limited numerical results show that the proposed PSO algorithm performs well for all test problems.

Quantum Particle Swarm Optimization (QPSO) is a global convergence guaranteed search method which introduces the Quantum theory into the basic Particle Swarm Optimization (PSO). QPSO performs better than normal PSO on several benchmark problems. However, QPSO’s quantum bit(Qubit) is still in Hilbert space’s unit circle with only one variable, so the quantum properties have been undermined to a large extent. In this paper, the Bloch Sphere encoding mechanism is adopted into QPSO, which can vividly describe the dynamic behavior of the quantum. In this way, the diversity of the swarm can be increased, and the local minima can be effectively avoided. The proposed algorithm, named Bloch QPSO (BQPSO), is tested with PID controller parameters optimization problem. Experimental results demonstrate that BQPSO has both stronger global search capability and faster convergence speed, and it is feasible and effective in solving some complex optimization problems.

This paper deals with the

m

-machine permutation flowshop scheduling problem to minimize the total flowtime, an NP-complete problem, and proposes an improved particle swarm optimization (PSO) algorithm. To enhance the exploitation ability of PSO, a stochastic iterated local search is incorporated. To improve the exploration ability of PSO, a population update method is applied to replace non-promising particles. In addition, a solution pool that stores elite solutions found in the search history is adopted, and in the evolution process each particle learns from this solution pool besides its personal best solution and the global best solution so as to improve the learning capability of the particles. Experimental results on benchmark instances show that the proposed PSO algorithm is competitive with other metaheuristics.

In this paper, a broadband MVDR(minimum variance distortionless response) beamforming method based on time-domain (TMVDR) is presented. Using TMVDR beamformer, stable sample matrix estimation could be obtained in short time period. To obtain the stable optimum solution of TMVDR, a numerical searching method optimized by PSO algorithm with constrain condition is introduced. Out-sea experiment show the performance of TMVDR beamformer applying PSO algorithm proposed in this paper.

The medical image registration algorithm uses the mutual information measure function that has many local extremes. Therefore, we propose our medical image registration algorithm that combines generalized mutual information with PSO-Powell hybrid algorithm and uses the objective measure function based on Renyi entropy. The Renyi entropy can remove the local extremes. We use the particle swarm optimization (PSO) algorithm to locate the measure function near the local extremes. Then we take the local extremes as initial point and use the Powell optimization algorithm to search for the global optimal solution. Section 2.2 of the paper presents the six-step procedure of our registration algorithm. We simulate medical image data with the registration algorithm; the simulation results, given in Table. 2 and 3, show preliminarily that the registration algorithm can eliminate the local extremes of objective measure function and accelerate the convergence rate, thus obtaining accurate and better registration results.

Relevance feedback (RF) is an iterative process which refines the retrievals by utilizing user’s feedback marked on retrieved results. Recent research has focused on the optimization for RF heuristic selection. In this paper, we propose an automatic RF heuristic selection framework which automatically chooses the best RF heuristic for the given query. The proposed method performs two learning tasks: query optimization and heuristic-selection optimization. The particle swarm optimization (PSO) paradigm is applied to assist the learning tasks. Experimental results tested on a content-based retrieval system with a real-world image database reveal that the proposed method outperforms several existing RF approaches using different techniques. The convergence behavior of the proposed method is empirically analyzed.

The purpose of the paper is to compare the performance of various fuzzy edge detectors which have been optimized by Particle Swarm Optimization (PSO). Three different type edge detectors Classical fuzzy Heuristic (CFH), Gaussian rule based (GRBF) and Robust Fuzzy Complement (RFC) are used. These edge detectors are effective in detecting edges, however the edges are thick. This paper proposes the used of particle swarm optimization algorithm as a method of producing thin and measurable edges. The fuzzy edge detectors are used in the initial swarm population and the objective function. The performance is based on the consistency of the visual appearance, fuzzy membership threshold and the number of complete edges detected. All three optimized edge detector performs reasonably well but CFHPSO outperform the rest.

One of the most studied problems in the financial is the intractability of the portfolios. Some practical formulations of the problem include various kinds of nonlinear constraints and objectives and can be efficiently solved by approximate algorithms. In this paper, we present a meta-heuristic algorithm named Particle Swarm Optimization (PSO) to the construction of optimal risky portfolios for financial investments. The PSO algorithm is tested on two portfolio optimization models and a comparative study with Genetic Algorithm has been implemented. The PSO model demonstrates high computational efficiency in constructing optimal risky portfolios. Preliminary results show that the approach is very promising and achieves results comparable or superior with the state of the art solvers.

This paper presents an optimal economic dispatch for power plants by using modified particle swarm optimization (PSO) algorithm. The economic dispatch problem in power systems is to determine the optimal combination of power outputs for all generating units in order that the total fuel cost can be minimized, furthermore, all practical constraints can be satisfied. Several key factors in terms of valve-point effects of coal cost functions, unit operation constraints and power balance are considered in the computation models. Consequently, a new adaptive PSO technique is utilized for solving economic dispatch problems. The proposed algorithm is compared with other PSO algorithms. Simulation results show that the proposed method is feasible and efficient.

This paper proposes a novel approach to optimal placement of wind turbines in the continuous space of a wind farm. The control objective is to maximize the power produced by a farm with a fixed number of turbines while guaranteeing the distance between turbines no less than the allowed minimal distance for turbine operation safety. The problem of wind farm micro-siting with space constraints is formulated to a constrained optimization problem and solved by a particle swarm optimization (PSO) algorithm based on penalty functions. Simulation results demonstrate that the PSO approach is more suitable and effective for micro-siting than the classical binary-coded genetic algorithms.

Table allocation is a type of assignment problem. The aim of table allocation is to assign multiple people to a single table in such a way that it minimizes a cost function. While particle swarm optimization (PSO) is normally used for continuous variables it has been adapted to solve this problem. Each particle represents an entire seating arrangement, and the velocity is the amount of times people swap tables during each iteration. In an example application PSO shows a significant improvement in fitness compared to the initial conditions, and has a low runtime. It also performs better in fitness improvement and runtime compared to choosing as many random samples as PSO generated. The use of PSO allows for generalized cost functions, and is simple to implement.

After the theorem which is used to determine whether all roots of a polynomial are in unit circle is given, and two particle swarm optimizations for finding the maximum module of the roots of a polynomial based on the theorem are proposed. Finally, several computer simulation results show that using these algorithms to find the maximum module of roots of a polynomial are more efficient and feasible, the convergent speed is much faster and the accuracy of results is much higher.

The ant colony algorithm (ACA) is a novel simulated evolutionary algorithm which is based on observations to behavior of some ant species. Because of the use of positive feedback mechanism, ACA has stronger robustness, better distributed computer system and easier to combine with other algorithms. However, it also has the flaws, for example mature and halting. This paper presents an optimization algorithm by used of multi-population hierarchy evolution. Each sub-population that is entrusted to different control achieves respectively a different search independently. Then, for the purpose of sharing information, the outstanding individuals are migrated regularly between the populations. The algorithm improves the parallelism and the ability of global optimization by the method. At the same time, according to the convex hull theory in geometry, the crossing point of the path is eliminated. Taking advantage of the common TSPLIB in international databases, lots of experiments are carried out. It is verified that the optimization algorithm effectively improves the convergence rate and the accuracy of reconciliation.

Parallel computing, network partitioning, and VLSI circuit placement are fundamental challenges in computer science. These problems can be modeled as graph partitioning problems. A new Similarity carrying Ant Model (SCAM) is used in the ant-based clustering algorithm to solve graph partitioning problem. In the proposed model, the ant will be able to collect similar items while it moves around. The flexible template mechanism had been used integrated with the proposed model to obtain the partitioning constrains. Random graph has been used to compare the new model with the original ant model and the model with short-term memory. The result of the experiments proves the impact of the SCAM compared with other models. This performance improvement for ant clustering algorithm makes it is feasible to be used in graph portioning problem.

Service-oriented grid environment enables a new way of service provisioning based on utility computing models, where users consume services based on their QoS (Quality of Service) requirements. In such “pay-per-use” Grids, workflow execution cost must be considered during scheduling based on users’ QoS constraints. In this paper, we propose a knowledge-based ant colony optimization algorithm (KBACO) for grid workflow scheduling with consideration of two QoS constraints, deadline and budget. The objective of this algorithm is to find a solution that minimizes execution cost while meeting the deadline in terms of users’ QoS requirements. Based on the characteristics of workflow scheduling, we define pheromone in terms of cost and design a heuristic in terms of latest start time of tasks in workflow applications. Moreover, a knowledge matrix is defined for the ACO approach to integrate the ACO model with knowledge model. Experimental results show that our algorithm achieves solutions effectively and efficiently.

Ant Colony Optimization (ACO) is recently proposed metaheuristic approach for solving hard combinatorial optimization problems. Parallel implementation of ACO can reduce the computational time obviously. An improved parallel ACO algorithm is proposed in this paper, which use dynamic transition probability to enlarge the search space by stimulating ants choosing new path at early stage; use polymorphic ant colony to improve convergence speed by local search and global search; use partially asynchronous parallel implementation, interactive multi-colony parallel and new information exchange strategy to improve the parallel efficiency. We implement the algorithm on the Dawn 4000L parallel computer using MPI and C language. The Numerical result indicates the algorithm proposed in this paper can improve convergence speed effectively with the fine solution quality.

In view of shortcomings of BP neural network, which is slow to converge and tends to trap in local optimum when applied in fault diagnosis, an approach for fault diagnosis based on BP neural network optimized by chaos ant colony algorithm is proposed. Mathematical model of chaos ant colony algorithm is created. Real-coded method is adopted and the weights and thresholds of BP neural network are taken as ant space position searched by chaos ant colony algorithm to train BP neural network. Training result of chaos ant colony algorithm is compared with that of conventional BP algorithm and from both results it is can be seen that chaos ant colony algorithm can overcome the shortcomings of BP algorithm. It is proved that mathematical model of chaos ant colony algorithm is correct and optimization method is valid through experimental simulation for machinery fault diagnosis of mine ventilator.

Laser range imaging is the current priority research areas of airborne lidar. And realizing accurate edge detection of laser range image is the key of completing the subsequent three-dimensional reconstruction. Based on the characteristics of laser range image and the deficiencies of traditional edge detection methods, a new improved fast adaptive ant colony algorithm for edge detection of laser range image has been proposed in this paper. Due to the initial cluster center and the heuristic guiding function used in the algorithm, the randomness and blindness of ants walking are eliminated thoroughly, and the speed of image edge detection is also greatly increased. Meanwhile, thanks to the applied ants’ selection mechanism and updating mechanism varying in contents, the error detection rate and omission factor of edge points as well as noise interference are all avoided, and the accuracy and adaptability of laser range image edge detection are greatly improved as well. Experimental results have shown that, this algorithm is more effective than other edge detection methods, and can meet the requirements of three-dimensional reconstruction.

A real-time moving ant estimator (RMAE) is developed for the bearings-only target tracking, in which ants located at their individual current state utilize the normalized weight and pheromone value to select the one-step prediction state and the dynamic moving velocity of each ant is depended directly on the normalized weights between two states. Besides this, two pheromone update strategy is implemented. Numerical simulations indicate that the RMAE could estimate adaptively the state of maneuvering or non-maneuvering target, and real-time requirement is superior to the moving ant estimator (MAE).

Facility layout planning plays an important role in the manufacturing process and seriously impacts a company’s profitability. A well-planned layout can significantly reduce the total material handling cost. The purpose of this paper is to develop a two-stage inter-cell layout optimization approach by using one of the popular meta-heuristics — the Ant Colony Optimization algorithm. At the first stage, the cells are formed based on the part-machine clustering results obtained through the ant system algorithm. In other words, we get the initial inter-cell layout after this stage. The work at the second stage uses a hybrid ant system algorithm to improve the solution obtained at previous stage. Different performance measures are also employed in this paper to evaluate the results.

A new boundary contour extraction algorithm based on curve evolution model and ant colony algorithm is proposed in this paper. Firstly, ant colony algorithm is used to find the optima of snake points for rapidly converging near image edge. Then the interpolation algorithm is applied to gaining the object’s rough contour that is used as the initial zero level set. The accurate contour can be obtained by the curve evolution method. Experimental results are given to demonstrate the feasibility of the proposed method in extracting contour from the blurred edge and high-noise images.

Ant Colony Optimization (ACO) is a heuristic bionic evolutive algorithm. In ACO algorithm, every ant has simple function, works with simple principle, which suits the characteristic of Wireless Sensor Networks (WSNs) and the request of its routing design. An ACO based Energy-Balance Routing Algorithm(ABEBR) was presented to balance the energy consumption in WSNs. Furthermore, a new pheromone update operator was designed to integrate energy consumption and hops into routing choice. This paper compares ABEBR with some classic routing algorithms (LEACH, DD and Flooding). Simulation results show that the presented algorithm can avoid energy working out too early on the less hops path, obviously balance the energy consumption and prolong the lifetime of WSNs.

Swarm Intelligence (SI) is a relatively new technology that takes its inspiration from the behavior of social insects and flocking animals. In this paper, we focus on two main SI algorithms: Ant Colony Optimization (ACO) and Particle Swarm Optimization (PSO). An extension of ACO algorithm and a PSO algorithm has been implemented to solve the portfolio optimization problem, which is a continuous multi-objective optimization problem.. The portfolio optimization model considered in this paper is based on the classical Markowitz mean-variance theory. The results show ACO performs better than PSO in the case of small-scale and large-scale portfolio, but in the case of medium-scale portfolio, PSO performs a better than ACO technique.

Automatic document classification is helpful in both organizing and finding information on huge resources. A novel multi-layered immune based document classification algorithm is presented. First, we represent the definition of the immune cells, antibody, antigen, and discuss the architecture of multi-layered immune system. Second, we evolve the dynamic models of immune response, immune regulation and immune memory, and establish the corresponding equations. Finally, we implement the simulation experiments, and compare the results with those obtained using the best methods for this application. Experiments show that the algorithm has higher classification accuracy than other document classification methods, and the attractive features such as diversity, self-learning, adaptive and robust etc. It provides a novel solution for document classification.

The study presents a novel quantum immune algorithm (QIA) for solving the parallel machine scheduling in the textile manufacturing industry. In this proposed algorithm, there are distinct characteristics as follows. First, the encoding method is based on Q-bit representation. Second, a novel mutation operator with a chaos-based rotation gate is proposed. Most importantly, two diversity schemes, suppression algorithm and similarity-based truncation algorithm, are employed to preserve the diversity of the population, and a new selection scheme is proposed to create the new population. Simulation results show that QIA is better than two quantum-inspired evolutionary algorithms.

A resource limited immune approach (RLIA) was developed to evolve architecture and initial connection weights of multilayer neural networks. Then, with Back-Propagation (BP) algorithm, the appropriate connection weights can be found. The RLIA-BP classifier, which is derived from hybrid algorithm mentioned above, is demonstrated on SPOT multi-spectral image data, vowel data and Iris data effectively. The simulation results demonstrate that RLIA-BP classifier possesses better performance comparing with Bayes maximum-likelihood classifier, k-nearest neighbor classifier (k-NN), BP neural network (BP-MLP) classifier and Resource limited artificial immune classifier (AIRS) in pattern classification.

In this paper, we present a new approach for the cryptanalysis of four-rounded Data Encryption Standard (DES) based on Artificial Immune System (AIS). The proposed algorithm is a combination of exploitation and exploration of fitness landscape where it performs local as well as global search. The algorithm has the property of automatically determining the population size and maintaining the local solutions in generations to generate results close to the global results. It is actually a known plaintext attack that aims at deducing optimum keys depending upon their fitness values. The set of deduced or optimum keys is scanned to extract the valuable bits out by counting all bits from the deduced key set. These valuable extracted bits produce a major divergence from other observed bits. This results in a 56-bit key deduction without probing the whole search space. To the best of our knowledge, the proposed algorithm is the first attempt to perform cryptanalysis of four-rounded DES using Artificial Immune System.

This paper proposes an immune concentration based virus detection approach which utilizes a two-element concentration vector to construct the feature. In this approach, ‘self’ and ‘nonself’ concentrations are extracted through ‘self’ and ‘nonself’ detector libraries, respectively, to form a vector with two elements of concentrations for characterizing the program efficiently and fast. Several classifiers including k-nearest neighbor (KNN), RBF neural network and support vector machine (SVM) with this vector as input are then employed to classify the programs. The selection of detector library determinant and parameters associated with a certain classifier is here considered as an optimization problem aiming at maximizing the accuracy of classification. A clonal particle swarm optimization (CPSO) algorithm is used for this purpose. Experimental results demonstrate that the proposed approach not only has a very much fast speed but also gives around 98% of accuracy under optimum conditions.

Inspired by observing fireworks explosion, a novel swarm intelligence algorithm, called Fireworks Algorithm (FA), is proposed for global optimization of complex functions. In the proposed FA, two types of explosion (search) processes are employed, and the mechanisms for keeping diversity of sparks are also well designed. In order to demonstrate the validation of the FA, a number of experiments were conducted on nine benchmark test functions to compare the FA with two variants of particle swarm optimization (PSO) algorithms, namely Standard PSO and Clonal PSO. It turns out from the results that the proposed FA clearly outperforms the two variants of the PSOs in both convergence speed and global solution accuracy.

Distribution network reconfiguration for loss minimization is a complex, large-scale combinatorial optimization problem. In this paper, a novel method called bacterial foraging optimization algorithm with particle swarm optimization strategy (BF-PSO) algorithm is applied to solve this problem. To verify the effectiveness of the proposed method, the optimization calculations of IEEE 69-bus testing system by the presented method are conducted and the calculation results are compared with pertinent literatures. Simulation results show that the proposed algorithm possesses fast convergence speed while the quality of solution and stability is ensured.

The article concern the influences of the novel flash structure with resistance wall on the forming load and die wear during the closed-die forging process. If the geometrical parameters of the resistance wall are not properly designed, the forming load and die wear may be great; this can affect the life of the forging die. Yet, no procedure is known to optimize and decide this flash structure at the exit. So, the optimization calculations were carried out using an authors optimal strategy, called Kriging-PSO strategy. According to this strategy we got the optimum parameters of the resistance wall. The strategy incorporates finite element simulations software Deform and includes particle swarm optimization (PSO) with a fast convergence and Kriging interpolation algorithm. The optimization results which represent the best die design were realized in industries.

Slab Stack Shuffling (SSS) problem is a kind of warehousing operations management problem abstracted from steel industry. SSS problem is to choose appropriate slabs for hot rolling schedule with the objective of minimizing shuffles during the retrieval process. Different from previous literatures, the substitute of slabs is a set of slabs which satisfy the given order demand. The problem in this paper considers balancing the shuffles between two sub-yards and the measurement of one shuffle is also different. The problem is formulated as an integer programming model by considering above practical requirements. The complexity of the model motivated us to develop a scatter search algorithm to solve the problem approximately. Problem-oriented coding scheme and solution combination method are proposed in scatter search. The computational results tested on real data show that the shuffles are decreased by 36.9% in average compared with the manual schedule.

To meet the requirement and solve the problems in system integration field, A Federation Structure based Multi-Agent System (FSMAS) model is proposed in this paper, with emphasis on the collaboration algorithm. This paper presents the process of partition and collaboration of the Agent tasks, the acquaintance first based on CNP algorithm in collaboration. FSMAS is applied to the development of agent-based system integration platform and tools. As a test case, a simulation system is developed which verifies the stability and efficiency of FSMAS in system integration filed.

GPU (Graphics Processing Unit) technology provides an efficient method for parallel computation. This paper will present a GPU-based Line Integral Convolution (LIC) parallel algorithm for visualization of discrete vector fields to accelerate LIC algorithm. The algorithm is implemented with parallel operations using Compute Unified Device Architecture (CUDA) programming model in GPU. The method can provide up to about 50× speed-up without any sacrifice on solution quality, compared to conventional sequential computation. Experiment results show that it is useful for in-time remote visualization of discrete vector fields.

Biogeography-based optimization algorithm(BBO) is a new kind of optimization algorithm based on Biogeography. It is designed based on the migration strategy of animals to solve the problem of optimization. In this paper, a new algorithm-Biogeography Migration Algorithm for Traveling Salesman Problem(TSPBMA) is presented. Migration operator is designed. It is tested on four classical TSP problems. The comparison results with the other nature inspired optimization algorithms show that TSPBMA is a very effective for TSP combination optimization. It provides a new way for this kinds of problem.

Redistricting is a Clustering Problem in optimization. The optimum redistricting is a convincing argument to voters that this solution is fair. In this paper, we set up a kind of model basing on the multi-factor model of clustering of the population pots by adopting the theory of optimization and the tools of stochastic simulation. Through this method, we solve the problem of how to realize the redistribution and the judging problem. Using the statistical data and practical model, we can get the districts of the state of New York satisfied with the rules of all principles.

Genetic algorithm is an evolutionary algorithm. It is particularly suitable for solving NP-complete optimization problems. In this paper, we propose a rapid genetic algorithm based on chaos mechanism. We introduce the chaos mechanism into the genetic algorithm. Using the ergodic property of the chaos movement, this method can remedy the defect of premature convergence in the genetic algorithm. In the search, this method continuously compresses the searching intervals of the optimization variable for increasing convergence speed. Experiments indicate that this method is a rapid and effective evolutionary algorithm.

The mathematics models of Reliability-based Structural Optimization (RBSO) were presented in this paper, then how to handle the constraint become sixty-four-dollar question of establishing the fitness function. Based on exterior penalty function method, penalty gene is made adaptively according to population’s evolution, then the fitness function is established, which is mapping formula of objective function and constraints. Subsequently laxity variable is introduced in primary mathematic model, based on Lagrange multiplier method, a new fitness function mapping formula is made, this method can avoid penalty function morbidity by means of adding a Lagrange multiplier, and has a more quick and stable convergence. Then, using GA successfully solved a numerical constrained optimization issue by this two mapping functions. The calculation shows that the two equations are reasonable and efficient, and Lagrange multiplier method has better global optimal capability.

This paper proposed a new algorithm of face recognition using genetic algorithm for classification. After initialization, selection and recombination are executed in a loop until fitness criterion is reached. The fitness function is calculated by an information theoretical measure. Experimental results presented here show that the proposed face recognition algorithm is effective.

An efficient search algorithm, which have to take into account the time and cost on emergency transport, the hierarchical Genetic Algorithm was proposed. Local optimization is achieved by Bottom GA in Subnet, global optimization is achieved by top GA in the whole network. The Contradiction of Global search capability and search efficiency is solved by maintaining a balance between GA Random search and this method is narrow search. The results of calculation show that this method can satisfy the demands of practical engineering of dynamic path selection in a wide range and quick fix of failure path on Emergency transport.

This paper proposed a new fault diagnosis method based on the extension genetic algorithm (EGA) for analog circuits. Analog circuits were difference at some node with the normal and failure conditions. However, the identification of the faulted location was not easily task due to the variability of circuit components. So this paper presented a novel EGA method for fault diagnosis of analog circuits, EGA is a combination of extension theory (ET) and genetic algorithm (GA). In the past, ET had to depend on experiences to set the classical domain and weight, but setting classical domain and weight were tedious and complicated steps in classified process. In order to improve this defect, this paper proposes an EGA to find the best parameter of classical domain and increase accuracy of the classification. The proposed method has been tested on a practical analog circuit, and compared with other classified method. The application of this new method to some testing cases has given promising results.

Collision detection is very important to enhance the sense of reality and immersion in virtual environment. Most of the traditional collision detection algorithms have been analyzed, but there is no algorithm that is applicable to all situations, and with the scene complexity increases, the efficiency of the algorithm tends to decline rapidly. In this paper, a new method is proposed to solve the problems: converting the problem of collision detection to the nonlinear programming problem with constraint conditions, and then using the adaptive genetic algorithm to solve it. The experiment results show that this method is efficient, especially in large-scale scenes.

Cooperative caching and relaying content in ISPs can decrease the bandwidth costs and distribution time. The relay resources installed at ISP are limited and the upload rates of relay servers are various. After formulating the optimization problem, we design a Non-dominated Sorting Bit matrix Genetic Algorithm (NSBGA) to solve it. Constraint-satisfied population is initialized according to resource ratio dynamically; improved alone point crossover and symmetric mutation is designed; population is non-dominated sorted. The experiments show that NSBGA is better than NSGAII and it can support P2P relay optimization very well. The relations between performances and parameters as the numbers of ISPs, source channels and relay servers are analyzed. As a general optimization algorithm, NSBGA also can be used in other application fields.

A novel parallel memetic algorithm (MA) architecture for the design of vector quantizers is presented in this paper. The architecture contains a number of modules operating memetic optimization concurrently. Each module uses steady-state genetic algorithm (GA) for global search, and K-means algorithm for local refinement. A shift register based circuit for accelerating mutation and crossover operations for steady state GA operations is adopted in the design. A pipeline architecture for the hardware implementation of K-means algorithm is also used. The proposed architecture is embedded in a softcore CPU, and implemented on a field programmable logic array (FPGA) device for physical performance measurement.

In order to solve traveling salesman problems that employed completion- time- shortest as the evaluating rule, an encoding method and improved differential evolution algorithm were proposed. In these methods, real number encoding and roulette wheel selection were adopted for improved differential evolution and neighborhood search operator was devised. It was fit for solving symmetric and asymmetric multiple traveling salesman problem. Asymmetric multiple traveling salesman problems were simulated. By comparison with the results of genetic algorithm and standard differential evolution, it is shown that the improved differential evolution algorithm proposed in this paper is efficient to solve the discrete combinatorial problem, such as optimization of multiple traveling salesman problems.

With rapid economic growth of Henan province, it is necessary to optimize all types of land resource allocation. Differential evolution algorithm is an evolutionary algorithm based on groups, which utilizes the differential information of individuals in the current population of solutions to guide its further search. There are three operators in Differential Evolution, mutation, crossover and selection. It applies that improved method to the optimization of land resource allocation. A model of land use allocation is put forward. The model can attain the optimal solution under multi-constraints such as economic benefit, coordinated and balanced program of development, the total population, agricultural acreage and environment. Results show that differential Evolution is effective and has robust character in dealing with multi-constraint and multi-dimensional optimization problems by cooperation and evolution of swarm.

A hybrid Differential Evolution algorithm with double population was proposed for 0-1 knapsack problem. The two populations play different roles during the process of evolution with the floating-point population as an engine while the binary population guiding the search direction. Each gene of every chromosome in the floating-point encoding population is restricted to the range [-1, 1], while each gene of every chromosome in the binary encoding population is zero or one. A new mapping operation based on sign function was proposed to generate the binary population from the floating-point population. And a local refining operation called discarding operation was employed in the new algorithm to fix up the solutions which are infeasible. Three benchmarks of 0-1 knapsack problem with different sizes were used to verify the new algorithm and the performance of the new algorithm was also compared with that of other evolutionary algorithms. The simulation results show it is an effective and efficient way for the 0-1 Knapsack problem.

In tree-based genetic programming (GP) performance optimization, the primary optimization target is the process of fitness evaluation. This is because fitness evaluation takes most of execution time in GP. Standard fitness evaluation uses the top-down tree evaluation algorithm. Top-down tree evaluation evaluates program tree from the root to the leaf of the tree. The algorithm reflects the nature of computer program execution and hence it is the most widely used tree evaluation algorithm. In this paper, we identify a scenario in tree evaluation where top-down evaluation is costly and less effective. We then propose a new tree evaluation algorithm called bottom-up tree evaluation explicitly addressing the problem identified. Both theoretical analysis and practical experiments are performed to compare the performance of bottom-up tree evaluation and top-down tree evaluation. It is found that bottom-up tree evaluation algorithm outperforms standard top-down tree evaluation when the program tree depth is small.

This paper examines the use of evolutionary computation (EC) to find optimal solution in vehicle assignment problem (VAP). The VAP refers to the allocation of the expected number of people in a potentially flooded area to various types of available vehicles in evacuation process. A novel discrete particle swarm optimization (DPSO) algorithm and genetic algorithm (GA) are presented to solve this problem. Both of these algorithms employed a discrete solution representation and incorporated a min-max approach for a random initialization of discrete particle position. A min-max approach is based on minimum capacity and maximum capacity of vehicles. We analyzed the performance of the algorithms using evacuation datasets. The quality of solutions were measured based on the objective function which is to find a maximum number of assigned people to vehicles in the potentially flooded areas and central processing unit (CPU) processing time of the algorithms. Overall, DPSO provides an optimal solutions and successfully achieved the objective function whereas GA gives sub optimal solution for the VAP.

The models of the evolution researchers are focused on the origin and the multiplication of organisms. This paper introduces a possible approach of the evolution of organisms being concerned in the appearance of the intelligence. This process is self developing and adaptive, producing the knowledge graph, which is the result of a lifelong data capturing and acquisition task of an artificial digital organism. This article tries to outline the appearance of the intelligence based on the principles of the creation process of a knowledge graph and its functions. The result is a non linear network of knowledge snippets, consisting of atoms, and their combinations, called contexts. Finally we introduce a startup information system, which is the realization of digital evolution machines and their ensemble in the artificial world. This special world is the world wide web.

In this paper, an improved thermodynamics evolutionary algorithm (ITEA) is proposed. The purpose of the new algorithm is to systematically harmonize the conflict between selective pressure and population diversity while searching for the optimal solutions. ITEA conforms to the principle of minimal free energy in simulating the competitive mechanism between energy and entropy in annealing process, in which population diversity is measured by similarity entropy and the minimum free energy is simulated with an efficient and effective competition by free energy component. Through solving some typical numerical optimization problems, satisfactory results were achieved, which showed that ITEA was a preferable algorithm to avoid the premature convergence effectively and reduce the cost in search to some extent.

Alopex is a correlation-based algorithm, which shares characteristics of both gradient descent approach and simulated annealing. It has been successfully applied to continuous and combinatorial optimization problems for years. Estimation of Distribution Algorithms (EDAs) is a class of novel evolutionary algorithms (EAs) proposed in recent years. Compared with the traditional EAs, it possesses unique evolutionary characteristics. In this paper, a hybrid evolutionary algorithm (EDA-Alopex) is proposed, which integrates the merits of both Alopex and EDA, and obtains more evolutionary information than these two approaches. The new algorithm is tested with several benchmark functions; numerical case study results demonstrate that EDA-Alopex outperforms both EDA and AEA, especially for the complex multi-modal functions. Finally, the proposed algorithm is investigated on high-dimensional and multi-peaks benchmark functions, and it also achieves satisfactory results.

By integrating artificial bee colony and genetic algorithm, a novel hybrid swarm intelligent approach is proposed in this paper. The main idea of the approach is to obtain the parallel computation merit of GA and the speed and self-improvement merits of ABC by sharing information between GA population and bee colony. To exam the proposed method, it is applied to 4 benchmark functions for different dimensions. For comparison, simple GA and ABC methods are also executed. Numerical results show that the proposed hybrid swarm intelligent method is effective, and the precision could be improved.

The job shop scheduling problem is a well-known NP hard problem, on which genetic algorithm is widely used. However, due to the lack of the major evolution direction, the effectiveness of the regular genetic algorithm is restricted. In this paper, we propose a new hybrid genetic algorithm to solve the job shop scheduling problem. The particle swarm optimization algorithm is introduced to get the initial population, and evolutionary genetic operations are proposed. We validate the new method on seven benchmark datasets, and the comparisons with some existing methods verify its effectiveness.

In this paper we construct an integer programming model for the order planning problem. Our model takes into account inventory matching and production planning simultaneously, and considers multiple objectives. We design a hybrid Particle Swarm Optimization, in which new heuristic rules to repair infeasible solutions are proposed, and then compare the results of using PSO, Tabu Search and the hybrid algorithm to solve the models of three different order quantities. Numerical results show that the hybrid PSO/TS algorithm provides more effective solutions.

In this work we propose a different particle swarm optimization (PSO) algorithm that employs two key features of the conjugate gradient (CG) method. Namely, adaptive weight factor for each particle and iteration number (calculated as in the CG approach), and periodic restart. Experimental results for four well known test problems have showed the superiority of the new PSO-CG approach, compared with the classical PSO algorithm, in terms of convergence speed and quality of obtained solutions.

The standard PSO has problems with consistently converging to good solutions, especially for multimodal functions. The reason for PSO failing to find (global) optima is premature convergence. Also, it has been shown in many empirical studies that PSO algorithms lack exploitation abilities. In this paper, we propose a hybrid of particle swarm optimization and local search, in which a standard PSO algorithm incorporates a local search algorithm. The standard PSO algorithm and the local search algorithm are devoted to exploration and exploitation of solution space, respectively. Particle’s current position is updated using update equation of standard PSO and then is refined by local search algorithm. The introduction of a local search improves the capability of exploitation of local region of standard PSO and prevents from premature convergence. The hybrid algorithm can locate multiple solutions without use of specific niching techniques. The hybrid algorithm showed superior performance on a set of multimodal functions.

The travelling salesman problem (TSP) is a classic problem of combinatorial optimization and is unlikely to find an efficient algorithm for solving TSPs directly. In the last two decades, ant colony optimization (ACO) has been successfully used to solve TSPs and their associated applicable problems. Despite the success, ACO algorithms have been facing constantly challenges for improving the slow convergence and avoiding stagnation at the local optima. In this paper, we propose a new hybrid algorithm, cooperative ant colony system and genetic algorithm (CoACSGA) to deal with these problems. Unlike the previous studies that regarded GA as a sequential part of the whole searching process and only used the result from GA as the input to the subsequent ACO iteration, this new approach combines both GA and ACS together in a cooperative and concurrent fashion to improve the performance of ACO for solving TSPs. The mutual information exchange between ACS and GA at the end of each iteration ensures the selection of the best solution for the next round, which accelerates the convergence. The cooperative approach also creates a better chance for reaching the global optimal solution because the independent running of GA will maintain a high level of diversity in producing next generation of solutions. Compared with the results of other algorithms, our simulation demonstrates that CoACSGA is superior to other ACO related algorithms in terms of convergence, quality of solution, and consistency of achieving the global optimal solution, particularly for small-size TSPs.

A controller of uncertain DC server motor is presented by using the fuzzy system with a real-time genetic algorithm. The parameters of the fuzzy system are online adjusted by the real-time genetic algorithm in order to generate appropriate control input. For the purpose of on-line evaluating the stability of the closed-loop system, an energy fitness function derived from backstepping technique is involved in the genetic algorithm. According to the experimental results, the genetic fuzzy control scheme performs on-line tracking successfully.

Multi-objective evolutionary algorithms (MOEAs) have been the mainstream to solve multi-objectives optimization problems. In this paper we add the static Bayesian game strategy into MOGA and propose a novel multi-objective genetic algorithm(SBG-MOGA). Conventional MOGAs use non-dominated sorting methods to push the population to move toward the real Pareto front. This approach has a good performance at earlier stages of the evolution, however it becomes hypodynamic at the later stages. In SBG-MOGA the objectives to be optimized are similar to players in a static Bayesian game. A player is a rational person who has his own strategy space. A player selects a strategy and takes an action to realize his strategy in order to achieve the maximal income for the objective he works on. The game strategy will generate a tensile force over the population and this will obtain a better multi-objective optimization performance. Moreover, the algorithm is verified by several simulation experiments and its performance is tested by different benchmark functions.

In this paper, we propose a Pareto-based tabu search algorithm for multi-objective FJSP with Earliness/Tardiness (E/T) penalty. In the hybrid algorithm, several neighboring structure based approaches were proposed to improve the convergence capability of the algorithm while keep population diversity of the last Pareto archive set. In addition, an external Pareto archive was developed to record the non-dominated solutions found so far. In the hybrid algorithm, dynamic parameters were introduced to adapt to the searching process. Experimental on several well-known benchmark instances show that the proposed algorithm is superior to several existing approaches in both solution quality and convergence ability.

The numerical simulation research of the Underwater Unmanned Vehicle (UUV) shell shape is carried out by applying the modern fluid dynamics numerical simulation technology. In the numerical simulation process, while focusing on the characteristics of the UUV shape, this treatise work properly with its computational model, computational domain, computational grid and boundary conditions. Based on numerical simulation techniques, the UUV shape is multi-objectively optimized by integrated fluid dynamics simulation software-fluent on iSIGHT optimization design platform. Multi-island genetic algorithm is adopted as the optimization algorithm, and Reynolds-averaged Navier-Stokes equation and turbulence model are as well applied in the optimization design process. The results show that by reducing the drag coefficient and the flow noise of the UUV shape, the optimized design has made a significant improvement in its comprehensive performance, which provides a new way for today’s optimization design of UUV shape.

Evacuation route planning is one of the most crucial tasks for solving massive evacuation problem. In large public places, pedestrians should be transferred to safe areas when nature or man-made accidents happen. A multi-objective ant colony algorithm for massive pedestrian evacuation is presented in this paper. In the algorithm, three objectives, total evacuation time of all evacuees, total routes risk degree and total crowding degree are minimized simultaneously. Ants search routes and converge toward the Pareto optimal solutions in the light of the pheromone. The experimental results show that the approach is efficient and effective to solve massive evacuation problem with rapid, reasonable and safe plans.

The study presents a novel weight-based multiobjective immune genetic algorithm(WBMOIGA), which is an improvement of its first version. In this proposed algorithm, there are distinct characteristics as follows. First, a randomly weighted sum of multiple objectives is used as a fitness function, and a local search procedure is utilized to facilitate the exploitation of the search space. Second, a new mate selection scheme, called tournament selection algorithm with similar individuals (TSASI), and a new environmental selection scheme, named truncation algorithm with similar individuals (TASI), are presented. Third, we also suggest a new selection scheme to create the new population based on TASI. Simulation results on three standard problems (ZDT3, VNT, and BNH) show WBMOIGA can find much better spread of solutions and better convergence near the true Pareto-optimal front compared to the elitist non-dominated sorting genetic algorithm (NSGA-II).

This paper proposes an enhanced mapping of multi-robot using a DSIFT to reduce the mapping calculation time. In this approach, the master robot transmits each robot’s mapping information in SLAM by DSIFT, which incorporates an additional step on the SIFT. The DSIFT uses a keypoint to reduce the distortional information throughout the Gaussian filter after the step of the image descriptor. The master robot calculates the slave robot’s pose using picture images, and serves the results to all the robots. Simulation results are presented based on DSIFT showing better performance than using the SIFT in multi-robot mapping situations.

A semiconductor fabrication line dynamic scheduling(SFLDS) model combining MAS(Multi-Agent System) with multi-intelligence algorithms is presented in this paper. The proposed model is based on the improved generalized partial global planning(GPGP) and utilizes the advantages of static intelligence algorithms with dynamic MAS. A scheduling process from ‘macro-scheduling to micro-scheduling to repeated- scheduling’ is designed for large-scale complex problems to enable to implement an effective and widely applicable prototype system for SFLDS. Under this scheme, a set of limitation and improvement of GPGP about its structure are proposed. The improved GPGP and its model are simulated by using simulation software eM—plant. A case study is provided to examine the practicability, flexibility and robustness of the proposed scheduling approach.

Formation is a good example of the research for multi-robot cooperation. Many different ways can be used to accomplish this task, but the main drawbacks of most of these methods are that robots can’t self-learn. In Brooks’ behavioral opinion, this paper is to verify that the reinforcement learning method can be used for robots to select different behaviors in various different situations. Experiments are performed to illustrate the team robots’ capability of self-learning and autonomy. The results show that the robots can get a self-formation in a barrier environment after learning.

This paper proposes a robust image stabilization system for a mobile robot using Extended Kalman Filter (EKF). Though image information is one of the most efficient data for robot navigation, it is subject to noise which results from internal vibration as well as external factors such as uneven terrain, stairs, or marshy surface. The vibration of camera deteriorates the definition of image by destroying image sharpness, which seriously prevents mobile robots from recognizing their environment for navigation. In this paper, inclinometer was used to measure the vibration angle of the camera system mounted on the robot to obtain a reliable image by compensating for the angle of the camera shake caused by vibration. In addition angle prediction by using the EKF enhances responsibility of image analysis for real time performance. The Experimental results show effectiveness of the proposed system to compensate for the blurring of the images.

This paper proposes an algorithm for an efficient navigation and building a precise map in multi-robot systems. One of the fundamental problems in mobile robotics is an effective investigation of unknown environments. The basis of navigation algorithm in this paper is Extented Wave Algorithm, which is in our point of view, appropriate in getting accurate.Secondly, particle filter, which proved its reliability, was considered as localization algorithm. Finally, overlapping algorithm is responsible for mapping. The technique has been tested extensively in simulation runs. The results given in this paper demonstrate that our algorithm significantly reduces the exploration time compared to previous approaches.

This present paper addresses impulsive consensus problem in directed networks of dynamic agents having communication delays. Based on impulsive control theory on delayed dynamical systems, a simple impulsive consensus protocol for such networks is proposed, and a generic criterion for solving average consensus problem is analytically derived. It is shown that global average consensus of a directed delayed networked multi-agent systems can be achieved by a suitable design of the impulsive gain and impulsive interval. Simulations are presented that are consistent with the theoretical results.

The basic theory of level of repair analysis (LORA) has been discussed. The route to apply multi-agent technology to accomplish the computer aided analysis for LORA has been investigated. A LORA system based on multi-agent system has been presented, the structure, the non-economic analysis agent has been researched. It can overcome the problem of information sharing and the cooperation between analysts effectively, and provide a new route to accomplish the computer aided analysis for LORA.

The basic content and procedure of Battlefield Damage Assessment(BDA) has been discussed and researched. The structure, the disposal strategy, the cooperation between agents, and the data flow of an Intelligent Battlefield Damage Assessment System(IBDAS) based on multi-agent system(MAS) has been studied. This system can solve the difficulty of BDA under the complicated and changing battlefield environment, and lay the theoretical foundation for the realization of a practical IBDAS based on multi-agent system.

In this paper, an adaptive system is proposed which attempts to combine together the approaches of studies of historical data and researches of multi-agent artificial market by evolving a double auction market model with diversity of different traders. The purpose of this research is to construct an artificial market which is more close to realistic one and more practical for future researches. The model with heterogeneous agents and the environment with which agents and market interact is complicated but controllable by data mining the optimal proportion of the different agents at the input to the market that generates an output which can fit historical data curve. The simulation results suggest that the system performance is close to the expecting values in the testing with adequate training in advance.

The average consensus in directed network of multi-agent with both switching topology and time-varying delay is studied. An orthogonal matrix is introduced to change the initial system into a reduced dimensional system. Based on linear matrix inequalities (LMIs) technique, a sufficient condition about average consensus problem is proposed. A novel form in terms of LMIs is obtained via taking the relationship between the terms in the Newton-Leibniz formula into account. Because some free weighted matrices are employed in the analysis processing and are selected through solving LMIs appropriately, our method is less conservative and more general. Finally, simulation examples are given to demonstrate the effectiveness of the theoretical results.

A virtual world is an online community in the form of a computer-based simulated environment, through which users can interact with one another and use and create objects. The non-player characters (NPC) in virtual world are following a fixed set of pre-programmed behaviors and lack the ability to adapt with the changing surrounding. Reinforcement learning agent is a way to deal with this problem. However, in a cooperative social environment, NPC should learn not only by trial and error, but also through cooperation by sharing information. The key investigation of this paper is: modeling the NPCs as multi-agent, and enable them to conduct cooperative learning, then speeding up the learning process. By using a fire fighting scenario in Robocup Rescue, our research shows that sharing information between cooperative agents will outperform independent agents who do not communicate during learning. The further work and some important issues of multi-agent reinforcement learning in virtual world will also be discussed in this paper.