ESSAWOA: Enhanced Whale Optimization Algorithm integrated with Salp Swarm Algorithm for global optimization
- PDF / 3,464,527 Bytes
- 18 Pages / 595.276 x 790.866 pts Page_size
- 113 Downloads / 275 Views
ORIGINAL ARTICLE
ESSAWOA: Enhanced Whale Optimization Algorithm integrated with Salp Swarm Algorithm for global optimization Qian Fan1 · Zhenjian Chen2 · Wei Zhang3 · Xuhua Fang1 Received: 20 September 2020 / Accepted: 28 September 2020 © Springer-Verlag London Ltd., part of Springer Nature 2020
Abstract In this paper, a novel hybrid meta-heuristic algorithm called ESSAWOA is proposed for solving global optimization problems. The main idea of ESSAWOA is to enhance Whale Optimization Algorithm (WOA) by combining the mechanism of Salp Swarm Algorithm (SSA) and Lens Opposition-based Learning strategy (LOBL). The hybridization process includes three parts: First, the leader mechanism with strong exploitation of SSA is applied to update the population position before the basic WOA operation. Second, the nonlinear parameter related to the convergence property in SSA is introduced to the two phases of encircling prey and bubble-net attacking in WOA. Third, LOBL strategy is used to increase the population diversity of the proposed optimizer. The hybrid design is expected to significantly enhance the exploitation and exploration capacity of the proposed algorithm. To investigate the effectiveness of ESSAWOA, twenty-three benchmark functions of different dimensions and three classical engineering design problems are performed. Furthermore, SSA, WOA and seven other well-known meta-heuristic algorithms are employed to compare with the proposed optimizer. Our results reveal that ESSAWOA can effectively and quickly obtain the promising solution of these optimization problems in the search space. The performance of ESSAWOA is significantly superior to the basic WOA, SSA and other meta-heuristic algorithms. Keywords Salp Swarm Algorithm · Whale Optimization Algorithm · Nonlinear parameter · Lens Opposition-based Learning · Hybridization
1 Introduction There are a large number of optimization problems in engineering applications and scientific research. It is essential to find the optimal solution to these problems under highly complex constraints in a reasonable time. Compared with conventional methods, meta-heuristic algorithms can usually obtain the optimal results on such problems. This is due to the fact that the algorithms have few parameters, can bypass the local optimum and do not require gradient information. Therefore, in recent years, meta-heuristic algorithms have attracted great attention and research in many fields. * Qian Fan [email protected] 1
College of Civil Engineering, Fuzhou University, Fuzhou 350116, China
2
School of Civil Engineering, Southeast University, Nanjing 210096, China
3
Fujian Academy of Building Research, Fuzhou 350025, China
Meta-heuristic algorithms can be generally divided into two categories: single-solution based metaheuristics and population-based metaheuristics [1]. Specifically, singlesolution based metaheuristics generate a set of candidate solutions according to the current single solution in the initial phase, and then replace the current solution with one of the
Data Loading...
