Showing papers on "Evolutionary programming published in 2012"

A Exploration and Exploitation in Evolutionary Algorithms: A Survey

Abstract: Every search algorithm needs to address the exploration and exploitation of a search space. Exploration is the process of visiting entirely new regions of a search space, whilst exploitation is the process of visiting those regions of a search space within the neighborhood of previously visited points. In order to be successful a search algorithm needs to establish a good ratio between exploration and exploitation. In this respect Evolutionary Algorithms (EAs) [De Jong 2002; Eiben and Smith 2008], such as Genetic Algorithms (GAs) [Michalewicz 1996; Goldberg 2008], Evolutionary Strategies (ES) [Back 1996], Evolutionary Programming (EP) [Fogel 1999], and Genetic Programming (GP) [Koza 1992], to name the more well-known instances, are no exception. Herrera and Lozano [1996] emphasized this by saying “The genetic algorithm behaviour is determined by the exploitation and exploration relationship kept throughout the run.” Many researchers believe that EAs are effective because of their good ratio between exploration and exploitation. Michalewicz [1996] stated that “Genetic Algorithms are a

Evolutionary dynamic optimization: A survey of the state of the art

Abstract: Optimization in dynamic environments is a challenging but important task since many real-world optimization problems are changing over time. Evolutionary computation and swarm intelligence are good tools to address optimization problems in dynamic environments due to their inspiration from natural self-organized systems and biological evolution, which have always been subject to changing environments. Evolutionary optimization in dynamic environments, or evolutionary dynamic optimization (EDO), has attracted a lot of research effort during the last 20 years, and has become one of the most active research areas in the field of evolutionary computation. In this paper we carry out an in-depth survey of the state-of-the-art of academic research in the field of EDO and other meta-heuristics in four areas: benchmark problems/generators, performance measures, algorithmic approaches, and theoretical studies. The purpose is to for the first time (i) provide detailed explanations of how current approaches work; (ii) review the strengths and weaknesses of each approach; (iii) discuss the current assumptions and coverage of existing EDO research; and (iv) identify current gaps, challenges and opportunities in EDO.

A Survey of Evolutionary Algorithms for Decision-Tree Induction

Abstract: This paper presents a survey of evolutionary algorithms that are designed for decision-tree induction. In this context, most of the paper focuses on approaches that evolve decision trees as an alternate heuristics to the traditional top-down divide-and-conquer approach. Additionally, we present some alternative methods that make use of evolutionary algorithms to improve particular components of decision-tree classifiers. The paper's original contributions are the following. First, it provides an up-to-date overview that is fully focused on evolutionary algorithms and decision trees and does not concentrate on any specific evolutionary approach. Second, it provides a taxonomy, which addresses works that evolve decision trees and works that design decision-tree components by the use of evolutionary algorithms. Finally, a number of references are provided that describe applications of evolutionary algorithms for decision-tree induction in different domains. At the end of this paper, we address some important issues and open questions that can be the subject of future research.

Encouraging behavioral diversity in evolutionary robotics: An empirical study

Abstract: Evolutionary robotics (ER) aims at automatically designing robots or controllers of robots without having to describe their inner workings. To reach this goal, ER researchers primarily employ phenotypes that can lead to an infinite number of robot behaviors and fitness functions that only reward the achievement of the task-and not how to achieve it. These choices make ER particularly prone to premature convergence. To tackle this problem, several papers recently proposed to explicitly encourage the diversity of the robot behaviors, rather than the diversity of the genotypes as in classic evolutionary optimization. Such an approach avoids the need to compute distances between structures and the pitfalls of the noninjectivity of the phenotype/behavior relation; however, it also introduces new questions: how to compare behavior? should this comparison be task specific? and what is the best way to encourage diversity in this context? In this paper, we review the main published approaches to behavioral diversity and benchmark them in a common framework. We compare each approach on three different tasks and two different genotypes. The results show that fostering behavioral diversity substantially improves the evolutionary process in the investigated experiments, regardless of genotype or task. Among the benchmarked approaches, multi-objective methods were the most efficient and the generic, Hamming-based, behavioral distance was at least as efficient as task specific behavioral metrics.

Swarm and Evolutionary Computation

Abstract: This paper presents a work inspired by the Pachycondyla apicalis ants behavior for the clustering problem. These ants have a simple but efficient prey search strategy: when they capture their prey, they return straight to their nest, drop off the prey and systematically return back to their original position. This behavior has already been applied to optimization, as the API meta-heuristic. API is a shortage of api-calis. Here, we combine API with the ability of ants to sort and cluster. We provide a comparison against Ant clustering Algorithm and K-Means using Machine Learning repository datasets. API introduces new concepts to ant-based models and gives us promising results.

Solution of multi-objective optimal power flow using gravitational search algorithm

Abstract: This article presents application of an efficient and reliable heuristic technique inspired by swarm behaviours in nature namely, gravitational search algorithm (GSA) for solution of multi-objective optimal power flow (OPF) problems. GSA is based on the Newton's law of gravity and mass interactions. In the proposed algorithm, the searcher agents are a collection of masses that interact with each other using laws of gravity and motion of Newton. In order to investigate the performance of the proposed scheme, multi-objective OPF problems are solved. A standard 26-bus and IEEE 118-bus systems with three different individual objectives, namely fuel cost minimisation, active power loss minimisation and voltage deviation minimisation, are considered. In multi-objective problem formulation fuel cost and loss; fuel cost and voltage deviation; fuel cost, loss and voltage deviation are minimised simultaneously. Results obtained by GSA are compared with mixed integer particle swarm optimisation, evolutionary programming, genetic algorithm and biogeography-based optimisation. The results show that the new GSA algorithm outperforms the other techniques in terms of convergence speed and global search ability.

Efficient constraint handling for optimal reactive power dispatch problems

Abstract: In power engineering, minimizing the power loss in the transmission lines and/or minimizing the voltage deviation at the load buses by controlling the reactive power is referred to as optimal reactive power dispatch (ORPD). Recently, the use of evolutionary algorithms (EAs) such as differential evolution (DE), particle swarm optimization (PSO), evolutionary programming (EP), and evolution strategies (ES) to solve ORPD is gaining more importance due to their effectiveness in handling the inequality constraints and discrete values compared to that of conventional gradient-based methods. EAs generally perform unconstrained searches, and they require some additional mechanism to handle constraints. In the literature, various constraint handling techniques have been proposed. However, to solve ORPD the penalty function approach has been commonly used, while the other constraint handling methods remain untested. In this paper, we evaluate the performance of different constraint handling methods such as superiority of feasible solutions (SF), self-adaptive penalty (SP), e -constraint (EC), stochastic ranking (SR), and the ensemble of constraint handling techniques (ECHT) on ORPD. The proposed methods have been tested on IEEE 30-bus, 57-bus, and 118-bus systems. Simulation results clearly demonstrate the importance of employing an efficient constraint handling method to solve the ORPD problem effectively.

Environment-driven distributed evolutionary adaptation in a population of autonomous robotic agents

Abstract: This article is concerned with a fixed-size population of autonomous agents facing unknown, possibly changing, environments. The motivation is to design an embodied evolutionary algorithm that can cope with the implicit fitness function hidden in the environment so as to provide adaptation in the long run at the level of population. The proposed algorithm, termed mEDEA, is shown to be both efficient in unknown environments and robust to abrupt and unpredicted changes in the environment. The emergence of consensus towards specific behavioural strategies is examined, with a particular focus on algorithmic stability. Finally, a real-world implementation of the algorithm is described with a population of 20 real-world e-puck robots.

Representations and operators for improving evolutionary software repair

Abstract: Evolutionary computation is a promising technique for automating time-consuming and expensive software maintenance tasks, including bug repair. The success of this approach, however, depends at least partially on the choice of representation, fitness function, and operators. Previous work on evolutionary software repair has employed different approaches, but they have not yet been evaluated in depth. This paper investigates representation and operator choices for source-level evolutionary program repair in the GenProg framework [17], focusing on: (1) representation of individual variants, (2) crossover design, (3) mutation operators, and (4) search space definition. We evaluate empirically on a dataset comprising 8 C programs totaling over 5.1 million lines of code and containing 105 reproducible, human-confirmed defects. Our results provide concrete suggestions for operator and representation design choices for evolutionary program repair. When augmented to incorporate these suggestions, GenProg repairs 5 additional bugs (60 vs. 55 out of 105), with a decrease in repair time of 17-43% for the more difficult repair searches.

Hybrid Metaheuristics Based on Evolutionary Algorithms and Simulated Annealing: Taxonomy, Comparison, and Synergy Test

Abstract: The design of hybrid metaheuristics with ideas taken from the simulated annealing and evolutionary algorithms fields is a fruitful research line. In this paper, we first present an overview of the hybrid metaheuristics based on simulated annealing and evolutionary algorithms presented in the literature and classify them according to two well-known taxonomies of hybrid methods. Second, we perform an empirical study comparing the behavior of a representative set of the hybrid approaches based on evolutionary algorithms and simulated annealing found in the literature. In addition, a study of the synergy relationships provided by these hybrid approaches is presented. Finally, we analyze the behavior of the best performing hybrid metaheuristic with regard to several state-of-the-art evolutionary algorithms for binary combinatorial problems. The experimental studies presented provide useful conclusions about the schemes for combining ideas from simulated annealing and evolutionary algorithms that may improve the performance of these kinds of approaches and suggest that these hybrids metaheuristics represent a competitive alternative for binary combinatorial problems.

On evolutionary computing in multi-ship trajectory planning

Abstract: The paper presents the updated version of Evolutionary Sets of Safe Ship Trajectories: a method which applies evolutionary algorithms and some of the assumptions of game theory to solving ship encounter situations. For given positions and motion parameters of the ships, the method finds a near optimal set of safe trajectories of all ships involved in an encounter. The method works in real time and the solutions must be returned within one minute, which enforces speeding up the optimization process. During the development of the method we have tested extensively various formulas for fitness function, problem-dedicated specialized operators as well as methods of selection. In the course of this research it turned out that some of the classic evolutionary mechanisms had to be modified for better performance, which included the order of some operations. The results of the adaptation process are presented here. The paper includes explicit description of all evolutionary mechanisms used and accentuates the research on improving the optimization process by adjusting evolutionary mechanisms to the problem.

Efficient aerodynamic design through evolutionary programming and support vector regression algorithms

Abstract: The shortening of the design cycle and the increase of the performance are nowadays the main challenges in aerodynamic design. The use of evolutionary algorithms (EAs) seems to be appropriate in a preliminary phase, due to their ability to broadly explore the design space and obtain global optima. Evolutionary algorithms have been hybridized with metamodels (or surrogate models) in several works published in the last years, in order to substitute expensive computational fluid dynamics (CFD) simulations. In this paper, an advanced approach for the aerodynamic optimization of aeronautical wing profiles is proposed, consisting of an evolutionary programming algorithm hybridized with a support vector regression algorithm (SVMr) as a metamodel. Specific issues as precision, dataset training size and feasibility of the complete approach are discussed and the potential of global optimization methods (enhanced by metamodels) to achieve innovative shapes that would not be achieved with traditional methods is assessed.

Evolutionary Computation for Reinforcement Learning

Abstract: Algorithms for evolutionary computation, which simulate the process of natural selection to solve optimization problems, are an effective tool for discovering high-performing reinforcement-learning policies. Because they can automatically find good representations, handle continuous action spaces, and cope with partial observability, evolutionary reinforcement-learning approaches have a strong empirical track record, sometimes significantly outperforming temporal-difference methods. This chapter surveys research on the application of evolutionary computation to reinforcement learning, overviewing methods for evolving neural-network topologies and weights, hybrid methods that also use temporal-difference methods, coevolutionary methods for multi-agent settings, generative and developmental systems, and methods for on-line evolutionary reinforcement learning.

Modeling of Biological Intelligence for SCM System Optimization

Abstract: This article summarizes some methods from biological intelligence for modeling and optimization of supply chain management (SCM) systems, including genetic algorithms, evolutionary programming, differential evolution, swarm intelligence, artificial immune, and other biological intelligence related methods. An SCM system is adaptive, dynamic, open self-organizing, which is maintained by flows of information, materials, goods, funds, and energy. Traditional methods for modeling and optimizing complex SCM systems require huge amounts of computing resources, and biological intelligence-based solutions can often provide valuable alternatives for efficiently solving problems. The paper summarizes the recent related methods for the design and optimization of SCM systems, which covers the most widely used genetic algorithms and other evolutionary algorithms.

Comparative study of evolutionary computation methods for active-reactive power dispatch

Abstract: This study presents a comparative study for four evolutionary computation (EC) methods to the optimal active–reactive power dispatch (ARPD) problem. Theoretically, there is a coupling relation between ARPDs. However, because of high X/R ratio existing in the transmission line, the problem of ARPD can be decomposed into two individual sub-problems by the decoupling concept, that is, ARPD problems. In this study, the evolutionary programming (EP), particle swarm optimisation (PSO), differential evolution (DE) and the proposed hybrid differential evolution (HDE) algorithms are used to separately solve the ARPD problem. To evaluate the performance of each method, the IEEE 30-bus and Taiwan Power Company (TPC) 345 kV simplified systems are employed as the study cases. The results indicate that the proposed HDE can obtain better results than the other methods in terms of active power transmission losses, voltage deviation, operating cost and convergence performance.

Evolving evolutionary algorithms

Abstract: This paper proposes a Grammatical Evolution framework to the automatic design of Evolutionary Algorithms. We define a grammar that has the ability to combine components regularly appearing in existing evolutionary algorithms, aiming to achieve novel and fully functional optimization methods. The problem of the Royal Road Functions is used to assess the capacity of the framework to evolve algorithms. Results show that the computational system is able to evolve simple evolutionary algorithms that can effectively solve Royal Road instances. Moreover, some unusual design solutions, competitive with standard approaches, are also proposed by the grammatical evolution framework.

Comparative Study on Distributed Generator Sizing Using Three Types of Particle Swarm Optimization

Abstract: Total power losses in a distribution network can be minimized by installing Distributed Generator (DG) with correct size. In line with this objective, most of the researchers have used multiple types of optimization technique to regulate the DG's output to compute its optimal size. In this paper, a comparative studies of a new proposed Rank Evolutionary Particle Swarm Optimization (REPSO) method with Evolutionary Particle Swarm Optimization (EPSO) and Traditional Particle Swarm Optimization (PSO) is conducted. Both REPSO and EPSO are using the concept of Evolutionary Programming (EP) in Particle Swarm Optimization (PSO) process. The implementation of EP in PSO allows the entire particles to move toward the optimal value faster. A test on determining optimum size of DGs in 69 bus radial distribution system reveals the superiority of REPSO over PSO and EPSO.

A Review on the Interpretability-Accuracy Trade-Off in Evolutionary Multi-Objective Fuzzy Systems (EMOFS)

Abstract: Interpretability and accuracy are two important features of fuzzy systems which are conflicting in their nature. One can be improved at the cost of the other and this situation is identified as “Interpretability-Accuracy Trade-Off”. To deal with this trade-off Multi-Objective Evolutionary Algorithms (MOEA) are frequently applied in the design of fuzzy systems. Several novel MOEA have been proposed and invented for this purpose, more specifically, Non-Dominated Sorting Genetic Algorithms (NSGA-II), Strength Pareto Evolutionary Algorithm 2 (SPEA2), Fuzzy Genetics-Based Machine Learning (FGBML), (2 + 2) Pareto Archived Evolutionary Strategy ((2 + 2) PAES), (2 + 2) Memetic- Pareto Archived Evolutionary Strategy ((2 + 2) M-PAES), etc. This paper introduces and reviews the approaches to the issue of developing fuzzy systems using Evolutionary Multi-Objective Optimization (EMO) algorithms considering ‘Interpretability-Accuracy Trade-off’ and mainly focusing on the work in the last decade. Different research issues and challenges are also discussed.

A direct local search mechanism for decomposition-based multi-objective evolutionary algorithms

Abstract: In recent years, the development of multi-objective evolutionary algorithms (MOEAs) hybridized with mathematical programming techniques has significantly increased. However, most of these hybrid approaches are gradient-based, and tend to require a high number of extra objective function evaluations to estimate the gradient information required. The use of direct search methods—i.e., methods that do not require gradient information—has been, however, less popular in the specialized literature (although such approaches have been used with single-objective evolutionary algorithms). This paper precisely focuses on the design of a hybrid between the wellknownMOEA/ D and Nelder and Mead's algorithm. Clearly, the mathematical programming technique adopted here, acts as a local search mechanism, whose goal is to improve the search performed by MOEA/D. Because of its nature, the proposed local search mechanism can be easily coupled to any other decomposition-based MOEA. Our preliminary results indicate that this sort of hybridization is quite promising for dealing with multi-objective optimization problems (MOPs) having high dimensionality (in decision variable space).

Diversity Guided Evolutionary Programming: A novel approach for continuous optimization

Abstract: Avoiding premature convergence to local optima and rapid convergence towards global optima has been the major concern with evolutionary systems research. In order to avoid premature convergence, sufficient amount of genetic diversity within the evolving population is considered necessary. Several studies have focused to devise techniques to control and preserve population diversity throughout the evolution. Since mutation is the major operator in many evolutionary systems, such as evolutionary programming and evolutionary strategies, a significant amount of research has also been done for the elegant control and adaptation of the mutation step size that is proper for traversing across the locally optimum points and reach for the global optima. This paper introduces Diversity Guided Evolutionary Programming, a novel approach to combine the best of both these research directions. This scheme incorporates diversity guided mutation, an innovative mutation scheme that guides the mutation step size using the population diversity information. It also takes some extra diversity preservative measures to maintain adequate amount of population diversity in order to assist the proposed mutation scheme. An extensive simulation has been done on a wide range of benchmark numeric optimization problems and the results have been compared with a number of recent evolutionary systems. Experimental results show that the performance of the proposed system is often better than most other algorithms in comparison on most of the problems.

Increasing Efficiency of Evolutionary Algorithms by Choosing between Auxiliary Fitness Functions with Reinforcement Learning

Abstract: In this paper further investigation of the previously proposed method of speeding up single-objective evolutionary algorithms is done. The method is based on reinforcement learning which is used to choose auxiliary fitness functions. The requirements for this method are formulated. The compliance of the method with these requirements is illustrated on model problems such as Royal Roads problem and H-IFF optimization problem. The experiments confirm that the method increases the efficiency of evolutionary algorithms.

A new approach to simultaneous localization and map building with implicit model learning using neuro evolutionary optimization

Abstract: This paper presents Neuro-Evolutionary Optimization SLAM (NeoSLAM) a novel approach to SLAM that uses a neural network (NN) to autonomously learn both a nonlinear motion model and the noise statistics of measurement data. The NN is trained using evolutionary optimization to learn the residual error of the motion model, which is then added to the odometry data to obtain the full motion model estimate. Stochastic optimization is used, to accommodate any kind of cost function. Prediction and correction are performed simultaneously within our neural framework, which implicitly integrates the motion and sensor models. An evolutionary programming (EP) algorithm is used to progressively refine the neural model until it generates a trajectory that is most consistent with the actual sensor measurements. During this learning process, NeoSLAM does not require any prior knowledge of motion or sensor models and shows consistently good performance regardless of the robot and the sensor noise type. Furthermore, NeoSLAM does not require the data association step at loop closing which is crucial in most other SLAM algorithms, but can still generate an accurate map. Experiments in various complex environments with widely-varying types of noise show that the learning capability of NeoSLAM ensures performance that is consistently less sensitive to noise and more accurate than that of other SLAM methods.