Marine Predators Algorithm: A nature-inspired metaheuristic

[koptimizer]

📋 논문의 정보를 알려주세요.

• Marine Predators Algorithm: A nature-inspired metaheuristic
• Afshin Faramarzi et al
• Expert Systems With Applications
• 2020-05-13

📃 Abstract

This paper presents a nature-inspired metaheuristic called Marine Predators Algorithm (MPA) and its application in engineering. The main inspiration of MPA is the widespread foraging strategy namely Lévy and Brownian movements in ocean predators along with optimal encounter rate policy in bio- logical interaction between predator and prey. MPA follows the rules that naturally govern in optimal foraging strategy and encounters rate policy between predator and prey in marine ecosystems. This paper evaluates the MPA’s performance on twenty-nine test functions, test suite of CEC-BC-2017, randomly generated landscape, three engineering benchmarks, and two real-world engineering de- sign problems in the areas of ventilation and building energy performance. MPA is compared with three classes of existing optimization methods, including (1) GA and PSO as the most well-studied metaheuristics, (2) GSA, CS and SSA as almost recently developed algorithms and (3) CMA-ES, SHADE and LSHADE-cnEpSin as high performance optimizers and winners of IEEE CEC competition. Among all methods, MPA gained the second rank and demonstrated very competitive results compared to LSHADE-cnEpSin as the best performing method and one of the winners of CEC 2017 competition. The statistical post hoc analysis revealed that MPA can be nominated as a high performance optimizer and is a significantly superior algorithm than GA, PSO, GSA, CS, SSA and CMA-ES while its perfor- mance is statistically similar to SHADE and LSHADE-cnEpSin. The source code is publicly available at: https://github.com/afshinfaramarzi/Marine-Predators-Algorithm, http://built-envi.com/portfolio/marine- predators-algorithm/, https://www.mathworks.com/matlabcentral/fileexchange/74578-marine-predators- algorithm-mpa, and http://www.alimirjalili.com/MPA.html.

🔎 어떤 논문인지 소개해주세요.

• 굉장히 따끈따끈한 SI 계열 메타휴리스틱 알고리즘입니다. IF도 높고, 코드도 공개되어 있어서 연구 가치가 높습니다.
• 레비 비행 뿐 아니라 브라운 운동까지 사용하는 것으로 보입니다. 매우 흥미롭습니다.

🔑 핵심 키워드를 적어주세요.

• Marine Predators Algorithm, Metaheuristic, Stochastic optimization, Evolutionary computation, Swarm intelligence

📎 URL

• https://www.sciencedirect.com/science/article/pii/S0957417420302025