General Evolutionary Algorithm Research Articles

A Particle Swarm Optimization-Based Generative Adversarial Network

At present, the combination of general evolutionary algorithms (EAs) and neural networks is limited to optimizing the framework or hyper parameters of neural networks. To further extend applications of EAs on neural networks, we propose a particle swarm optimization (PSO) based generative adversarial network(GAN), named as PGAN in this paper. In the study, PSO is utilized as a generator to generate fake data, while the discriminator is a traditional fully connected neural network. In the confrontation process, when the proposed PSO can generate a better fake image, this will react to the discriminator, so that the discriminator can improve the recognition effect of the image and the better discriminator also accelerates the evolution of the overall model. Through experiments, we explore the new application value of EAs in deep learning, so that the sample data in EAs and the sample data in deep learning are interconnected. The PSO algorithm is improved, so that it truly participates in the confrontation with multi-layer perceptrons.

Analysis of the model using General Evolutionary Algorithm

The study aims to solve the universality of evolutionary computation for solving different optimization problems, avoid the repeating design, and share codes, so as to improve the application efficiency of the algorithms. To achieve this, the authors put forward a general evolutionary model based on object-oriented programming language (C# language) in this paper. This model divides the problem domain, the fitness function, the algorithm and the parameter control into different classes, which are encapsulated into different objects so as to enhance the independence of each module. By utilizing the model, it is convenient to develop, reuse, extend and modify software while guaranteeing the universality of the algorithm at the same time.

Evolutionary and Ideation Concepts for Cybersecurity Education

Evolution is a well-established biological theory, but some basic concepts can be abstracted and applied to non-biological domains such as the education domain for the purposes of knowledge sharing. There is a gap in the literature regarding how evolutionary processes can be applied to cyber security education. This article presents the general evolutionary algorithm and pairs it with an ideation technique (SCAMPER) to illustrate how certain evolutionary processes can be applied to cyber security education and learning. This paper does not attempt to close the gap, but rather offer a theoretical approach to address the gap.

ImGA: an improved genetic algorithm for partitioned scheduling on heterogeneous multi-core systems

Efficient mapping of tasks onto heterogeneous multi-core systems is very challenging especially in the context of real-time applications. Assigning tasks to cores is an NP-hard problem and solving it requires the use of meta-heuristics. Relevantly, genetic algorithms have already proven to be one of the most powerful and widely used stochastic tools to solve this problem. Conventional genetic algorithms were initially defined as a general evolutionary algorithm based on blind operators with pseudo-random operations. It is commonly admitted that the use of these operators is quite poor for an efficient exploration of big problems. Likewise, since exhaustive exploration of the solution space is unrealistic, a potent option is often to guide the exploration process by hints, derived by problem structure. This guided exploration prioritizes fitter solutions to be part of next generations and avoids exploring unpromising configurations by transmitting a set of predefined criteria from parents to children. Consequently, genetic operators, such as initial population, crossover, mutation must incorporate specific domain knowledge to intelligently guide the exploration of the design space. In this paper, an improved genetic algorithm (ImGA) is proposed to enhance the conventional implementation of this evolutionary algorithm. In our experiments, we proved that ImGA leads to perceptible increase in the performance of the genetic algorithm and its convergence capabilities.

Convergence analysis of general evolutionary algorithms

Traditional Evolutionary Algorithm(EA) convergence research focuses on specific algorithm;consequently the conclusion is only suitable for some specific algorithm.In order to study the convergence of all EAs,this paper presented a general EA including EAs of all operator types.A probability space was set up for the purpose of studying the algorithm's convergence,and all terms on the algorithm were strictly defined in mathematical language,and seven theorems related to the algorithm's convergence were completely proved in the probability space.One of the theorems found the sufficient and necessary conditions for the algorithm's convergence in probability.More importantly,these theorems are suitable to all types of EAs.A system composed of these theorems was established,which could be used to guide the EA design and judge the correctness of an EA theoretically.

An incremental evolutionary learning method for optimizing content-based image indexing algorithms

One of the future directions of content-based image retrieval (CBIR) systems is incremental learning of indexing and retrieval algorithms. Optimization of the indexing algorithm is more difficult compared to the retrieval algorithm enhancement; since each time the indexing algorithm parameters are modified, all images of the reference database should be indexed again. This paper considers, for the first time, a challengeable limitation of actual indexing optimization systems: learning in dynamic and incremental CBIR environments. We introduce a new incremental evolutionary optimization method based on evolutionary group algorithm. The new incremental evolutionary group algorithm (IEGA) overcomes time-consuming drawbacks related to general evolutionary algorithms in large scale content-based image indexing optimization tasks and presents a new strategy that is enhanced with the ability of incremental learning. Evaluation results on some simulated dynamic CBIR systems show that the proposed method can continuously obtain good performance in the presence of environmental or scale changes.

Multiobjective optimization of MPLS-IP networks with a variable neighborhood genetic algorithm

This paper presents a Genetic Algorithm for the optimization of multiple indices of Quality of Service of Multi Protocol Label Switching (MPLS) IP networks. The proposed algorithm, the Variable Neighborhood Multiobjective Genetic Algorithm (VN-MGA), is a Genetic Algorithm based on the NSGA-II, with the particular feature that solutions are encoded defining two different kinds of neighborhoods. The first neighborhood is defined by considering as decision variables the edges that form the routes to be followed by each request, whilst the second part of solution is kept constant. The second neighborhood is defined by considering the request sequence as decision variable, with the first part kept constant. Comparisons are performed with: (i) a VNS algorithm that performs a switch between the same two neighborhoods that are used in VN-MGA; and (ii) the results obtained with an integer linear programming solver, running a scalarized version of the multiobjective problem. The results indicate that the proposed VN-MGA outperforms the pure VNS algorithm, and provides a good approximation of the exact Pareto fronts obtained with Integer Linear Programming (ILP) approach, at a much smaller computational cost. Besides potential benefits of the application of the proposed approach to the optimization of packet routing in MPLS networks, this work raises the theoretical issue of the systematic application of variable encodings, which allow variable neighborhood searches, as generic operators inside general evolutionary computation algorithms.

Theory and applications of swarm intelligence

Swarm intelligence is the discipline that deals with natural and artificial systems composed of many individuals that coordinate their activities using decentralized control and self-organization. In particular, the discipline focuses on the behavior of social insects such as fish schools and bird flocks and colonies of ants, termites, bees, and wasps. The most well-known examples of systems studied by swarm intelligence are particle swarm optimization (PSO) and ant colony optimization (ACO). Particle swarm optimization mimics the behavior of fish schooling and bird flocking. PSO is a population-based stochastic optimization strategy with fast convergent speed than general evolutionary algorithms (EAs). Different from EAs, each particle employs not only position information but also the velocity information. They communicate good positions to each other and adjust their own positions according to their decision. In PSO, a number of simple entities—the particles—are placed in the search space of some problem or function, and each evaluates the objective function at its current location. Each particle then determines its movement through the search space by combining some aspect of the history of its own current and best (best-fitness) locations with those of one or more members of the swarm, with some random perturbations. The next iteration takes place after all particles have been moved. Eventually, the swarm as a whole, like a flock of birds collectively foraging for food, moves close to an optimum of the fitness function. Ant colony optimization is a probabilistic technique for solving computational hard problems which can be reduced to finding optimal paths. The main idea is the indirect communication between the ants by means of chemical pheromone trails. The special issue on the swarm intelligence is dedicated to the latest work in the theory and applications in this exciting area. Our aim is to provide a useful reference for understanding new trends on swarm intelligence. After a detailed review process, a total of four papers were selected to reflect the call thematic vision. The contents of these studies are briefly described as follows. The first paper entitled ‘Analysis of emergent symmetry breaking in collective decision making’ by Heiko Hamann, Thomas Schmickl, Heinz Worn, and Karl Crailsheim investigates a simulated multi-agent system (MAS) that collectively decides to aggregate at an area of high utility. The agents’ control algorithm is based on random agent– agent encounters and is inspired by the aggregation behavior of honey bees. In this article, symmetry breaking is defined, several symmetry-breaking measures are proposed, and phenomenon of emergent symmetry breaking within authors’ observed system is reported. The ability of the MAS to successfully break the symmetry depends significantly on a local-neighborhood-based threshold of the agents’ control algorithm that determines at which number of neighbors the agents stop. This dependency is analyzed, and two macroscopic features are determined that significantly influence the symmetry-breaking behavior. In addition, the connection between the ability of the Z. Cui (&) Complex System and Computational Intelligence Laboratory, Taiyuan University of Science and Technology, 030024 Shanxi, Taiyuan, China e-mail: cuizhihua@gmail.com

Coupled Genetic Algorithm—Linear Programming Scheme for Least-Cost Pipe Sizing of Water-Distribution Systems

Water-distribution systems least-cost pipe sizing/design is probably the most explored problem in water-distribution systems optimization. Attracted numerous studies over the last 4 decades, two main approaches were employed: decomposition in which an “inner” linear programming problem is solved for a fixed set of flows/heads, while the flows/heads are altered at an “outer” problem using a gradient or a subgradient type technique; and the employment of a general evolutionary optimization algorithm. In 1995 Loganathan and his colleagues proposed to couple these two approaches into one framework, thus overcoming the limitations of each. This study employs this framework with two modifications: (1) application of a genetic algorithm for the “outer” optimization search instead of simulated annealing; and (2) constraining the sought solution to the lowest cost spanning tree layout with the spanning tree chords kept at their minimum permissible pipe diameters. A comparison of the methodology to a genetic algori...