• PDF / 2,798,949 Bytes
• 26 Pages / 595.276 x 790.866 pts Page_size
• 66 Downloads / 233 Views

DOWNLOAD

REPORT

---

RESEARCH ARTICLE-COMPUTER ENGINEERING AND COMPUTER SCIENCE

Modified Harris Hawks Optimization Algorithm for Global Optimization Problems Yang Zhang1 · Xizhao Zhou1 · Po-Chou Shih2 Received: 5 December 2019 / Accepted: 18 August 2020 © King Fahd University of Petroleum & Minerals 2020

Abstract The Harris hawks optimization algorithm (HHO) is a novel swarm-based meta-heuristic algorithm. In this study, a modified Harris hawks optimization algorithm (MHHO) is proposed to enhance the searching performance of the conventional HHO. Past studies have revealed that different adjustment strategies of important variables in meta-heuristic algorithm will evidently affect the performance of the algorithm. Therefore, this study focuses on the escaping energy (E) of prey is an extremely, which is a critical concept that determines the balance between the exploration and exploitation phases of the HHO. In nature, the Harris hawks will take different the perch strategy and the chasing pattern according to E. For E, six different update strategies are designed to model the real situation. To explore the differences between the six strategies mentioned above, a comparative study through twenty representative benchmark functions is carried out by Experiment 1 (Sect. 4.2). The results show that strategy 6 (the exponential decreasing strategy) outperforms other rivals; therefore, it is deployed into the MHHO. To further demonstrate the superior search performance of MHHO, a similar comparative study between MHHO and several well-established optimization technologies is carried out by Experiment 2 (Sect. 4.3). The results clearly exhibit MHHO outperforms its rivals in most benchmark functions. In addition, compared with other well-known optimizers and the conventional HHO, the competitive results obtained by MHHO on two engineering optimization problems also prove the effectiveness and superiority of the proposed MHHO in solving constrained optimization problems. Keywords Meta-heuristic · Global optimization · Harris hawks optimization algorithm · Evolutionary computation

1 Introduction A meta-heuristic algorithm (MHA) is a technique that obtains approximate optimality under certain time and occasion constraints, which are favored for their simplicity, efficiency, and low computational cost. The study of MHA began in the 1980s; in 1983, Kirkpatrick et al. proposed the simulated annealing algorithm (SA) [1] by combining the concepts of thermodynamic annealing and the Monte Carlo algorithm.

B

Yang Zhang [email protected] Xizhao Zhou [email protected] Po-Chou Shih [email protected]

1

Business School, University of Shanghai for Science and Technology, Shanghai 200093, China

2

Department of Industrial and Management, Chaoyang University of Technology, Taichung 413310, Taiwan

In 1992, the genetic algorithm (GA) [2] based on Darwin’s theory of evolution was proposed by Holland. Then, the particle swarm optimization algorithm (PSO) [3] was proposed by Kennedy et al. in 1995 and the ant colony optimization algorithm (ACO) [4] was des