A novel mutation strategy selection mechanism for differential evolution based on local fitness landscape
- PDF / 1,233,991 Bytes
- 31 Pages / 439.37 x 666.142 pts Page_size
- 62 Downloads / 229 Views
A novel mutation strategy selection mechanism for differential evolution based on local fitness landscape Zhiping Tan1 · Kangshun Li1 · Yuan Tian2 · Najla Al‑Nabhan3 Accepted: 20 October 2020 © Springer Science+Business Media, LLC, part of Springer Nature 2020
Abstract The performance of differential evolution (DE) algorithm highly depends on the selection of mutation strategy. However, there are six commonly used mutation strategies in DE. Therefore, it is a challenging task to choose an appropriate mutation strategy for a specific optimization problem. For a better tackle this problem, in this paper, a novel DE algorithm based on local fitness landscape called LFLDE is proposed, in which the local fitness landscape information of the problem is investigated to guide the selection of the mutation strategy for each given problem at each generation. In addition, a novel control parameter adaptive mechanism is used to improve the proposed algorithm. In the experiments, a total of 29 test functions originated from CEC2017 single-objective test function suite which are utilized to evaluate the performance of the proposed algorithm. The Wilcoxon rank-sum test and Friedman rank test results reveal that the performance of the proposed algorithm is better than the other five representative DE algorithms. Keywords Adaptive mutation strategy · Local fitness landscape · Differential evolution · Parameter adaptation
1 Introduction Optimization is an important technique in real-world life, which is widely used in engineering domain, agriculture and industry [1]. In general, an unconstrained optimization problem is identified as a single-objective global optimization which * Kangshun Li [email protected] Zhiping Tan [email protected] 1
College of Mathematics and Informatics, South China Agricultural University, Guangzhou 510642, China
2
School of Computer Engineering, Nanjing Institute of Technology, Nanjing 210024, China
3
Computer Science, King Saud University, Riyadh, Kingdom of Saudi Arabia
13
Vol.:(0123456789)
Z. Tan et al.
can be described as follows [2]. We need to search for a decision variables of { } vector x⃗ = x1 , x2 , ..., xD , which satisfies the(variable bound, xj,low ≤ x ≤ xj,upp and ) minimizes or maximizes a fitness function f x⃗ , where D denotes the dimension of the solution space, xj,low and xj,upp is the lower and upper bounds, respectively, j = 1,2,…,D. In practical problems, the decision variable can be discrete, continuous or hybrid and the objective function can be uni-modal or multi-modal, hybrid and composition. Therefore, the single-objective global optimization problems have various different characteristics and complicated mathematical proprieties, which make the algorithms difficult to get the desired results [3]. As exhaustive search-based methods are time expensive when solving complex optimization problems. Evolutionary algorithms (EAs) have been widely used in recently years [4]. EAs are population-based search methods that have been proved to be a promising approach t
Data Loading...
