Cooperative co-evolution has been introduced into evolutionary algorithms with the aim of solving increasingly complex optimization problems through a divide-and-conquer paradigm. In theory, idea co-adapted subcomponents is desirable for large-scale problems. However, in practice, without prior knowledge about problem, it not clear how problem should be decomposed. this paper, we propose an automatic decomposition strategy called differential grouping that can uncover underlying interaction...

Identification of variable interaction is essential for an efficient implementation a divide-and-conquer algorithm large-scale black-box optimization. In this paper, we propose improved variant the differential grouping (DG) algorithm, which has better efficiency and accuracy. The proposed DG2, finds reliable threshold value by estimating magnitude roundoff errors. With respect to efficiency, DG2 reuses sample points that are generated detecting interactions saves up half computational...

Many evolutionary algorithms have been proposed for large scale optimization. Parameter interaction in non-separable problems is a major source of performance loss specially on problems. Cooperative Co-evolution(CC) has as natural solution optimization problems, but lack systematic way decomposing obstacle CC frameworks. The aim this paper to propose capturing interacting variables more effective problem decomposition suitable cooperative co-evolutionary Grouping different subcomponents...

This article proposes a competitive divide-and-conquer algorithm for solving large-scale black-box optimization problems which there are thousands of decision variables and the algebraic models unavailable. We focus on that partially additively separable, since this type problem can be further decomposed into number smaller independent subproblems. The proposed addresses two important issues in optimization: (1) identification subproblems without explicitly knowing formula objective function...

In this paper we propose three techniques to improve the performance of one major algorithms for large scale continuous global function optimization. Multilevel Cooperative Co-evolution (MLCC) is based on a Co-evolutionary framework and employs technique called random grouping in order group interacting variables subcomponent. It also uses another adaptive weighting co-adaptation subcomponents. We prove that probability subcomponent using drops significantly as number increases. This calls...

Cooperative co-evolution (CC) is an explicit means of problem decomposition in multipopulation evolutionary algorithms for solving large-scale optimization problems. For CC, subpopulations representing subcomponents a co-evolve, and are likely to have different contributions the improvement best overall solution problem. Hence, it makes sense that more computational resources should be allocated with greater contributions. In this paper, we study how allocate context subsequently propose new...

Standard Cooperative Co-evolution uses a round-robin method to select subcomponents undergo optimization. In non-separable (epistatic) optimization problem, dividing the computational budget equally between all of is not necessarily best strategy. When dealing with problems, there usually an imbalance contribution various global fitness individuals. Using fashion treats and wastes budget. this paper, we propose Contribution Based (CBCC) that selects based on their contributions fitness. This...

Scalability is a crucial aspect of designing efficient algorithms. Despite their prevalence, large-scale dynamic optimization problems are not well studied in the literature. This paper concerned with benchmarks and frameworks for study problems. We start by formal analysis moving peaks benchmark (MPB) show its nonseparable nature irrespective number peaks. then propose composite MPB suite exploitable modularity covering wide range scalable partially separable functions suitable The exhibits...

In this paper we use a divide-and-conquer approach to tackle large-scale optimization problems with overlapping components. Decomposition for an problem is challenging as its components depend on one another. The existing decomposition methods typically assign all the linked decision variables into group, thus cannot reduce original size. To address issue modify Recursive Differential Grouping (RDG) method decompose problems, by breaking linkage at shared multiple evaluate efficacy of our...

In this paper, we propose a metric for evaluating the performance of user-preference based evolutionary multiobjective algorithms by defining preferred region on location user-supplied reference point. This uses composite front which is type set and used as replacement Pareto-optimal front. constructed extracting non-dominated solutions from merged solution sets all that are to be compared. A then defined Once defined, existing multi-objective metrics can applied with respect region. paper...

In this paper we investigate the performance of cooperative co-evolutionary (CC) algorithms on large-scale fully-separable continuous optimization problems. We have shown that decomposition can significant impact CC algorithms. The empirical results show subcomponent size should be chosen small enough so is within capacity optimizer. practice, determining optimal difficult. Therefore, adaptive techniques are desired by practitioners. Here propose an method, MLSoft, uses widely-used in...

Cooperative Co-evolution (CC) is a promising framework for solving large-scale optimization problems. However, the round-robin strategy of CC not an efficient way allocating available computational resources to components imbalanced functions. The imbalance problem happens when partially separable function have non-uniform contributions overall objective value. Contribution-Based (CBCC) variant that allocates individual based on their contributions. CBCC variants (CBCC1 and CBCC2) shown...

Problem decomposition plays an essential role in the success of cooperative co-evolution (CC), when used for solving large-scale optimization problems. The recently proposed recursive differential grouping (RDG) method has been shown to be very efficient, especially terms time complexity. However, it requires appropriate parameter setting estimate a threshold value order determine if two subsets decision variables interact or not. Furthermore, using one global may insufficient identify...

In this paper we propose a user-preference based evolutionary algorithm that relies on decomposition strategies to convert multi-objective problem into set of single-objective problems. The use reference point allows the focus search more preferred regions which can potentially save considerable amount computational resources. proposed, dynamically adapts weight vectors and is able converge close regions. Combining with approaches paves way for effective optimization many-objective method...

Evolutionary algorithms that rely on dominance ranking often suffer from a low selection pressure problem when dealing with many-objective problems. Decomposition and user-preference based methods can help to alleviate this great extent. In paper, evolutionary multi-objective algorithm is proposed uses decomposition for solving techniques are widely used in optimization require set of evenly distributed weight vectors generate diverse solutions the Pareto-optimal front. The newly algorithm,...

Dynamic changes are an important and inescapable aspect of many real-world optimization problems. Designing algorithms to find track desirable solutions while facing challenges dynamic problems is active research topic in the field swarm evolutionary computation. To evaluate compare performance algorithms, it imperative use a suitable benchmark that generates problem instances with different controllable characteristics. In this article, we give comprehensive review existing benchmarks...

Dynamic optimization problems (DOPs) are that change over time. However, most investigations in this domain focused on tracking moving optima (TMO) without considering the cost of switching from one solution to another when environment changes. Robust time (ROOT) tries address shortcoming by finding solutions which remain acceptable for several environments. ROOT methods only they become unacceptable. Indeed, TMO and two extreme cases sense former, is considered zero latter, it very large....

Many real-world optimization problems are dynamic. The field of robust over time (ROOT) deals with dynamic in which frequent changes the deployed solution undesirable. This can be due to high cost switching solutions, limitation needed resources deploy such new and/or system being intolerant towards solution. In considered ROOT this article, main goal is find solutions that maximize average number environments where they remain acceptable. state-of-the-art methods developed tackle these...