New binary marine predators optimization algorithms for 0–1 knapsack problems. (January 2021)

• Record Type: Journal Article
• Authors: Abdel-Basset, Mohamed; Mohamed, Reda; Chakrabortty, Ripon K.; Ryan, Michael; Mirjalili, Seyedali
• Is Part Of: Computers & industrial engineering. Volume 151(2021)
• Publication Date: 2021-01
• Subjects: 0–1 knapsack problem -- Marine predators algorithm -- Algorithm -- Artificial Intelligence -- Repair infeasible solutions -- Binary optimization -- Benchmark -- Combinatorial optimization; Engineering -- Data processing -- Periodicals; Industrial engineering -- Periodicals; 620.00285
• DOI: 10.1016/j.cie.2020.106949
• Languages: English
• ISSNs: 0360-8352

Abstract

Highlights: A New binary Marine Predators Optimization Algorithm is developed for 0–1 KPs. An initial elite population generator is proposed. We used large size instances to test the proposed algorithm. The proposed algorithm is compared with recent state-of-the-art algorithms. The superior performance of the proposed algorithm is proved in the experiments. Abstract: Recently, a novel meta-heuristic algorithm known as the marine predators algorithm (MPA) has been proposed for solving continuous optimization problems. Despite the significant superiority of MPA in solving continuous problems, the algorithm is not applicable to binary problems. This work proposes a binary version of MPA for solving the 0–1 knapsack (KP01) problem. To develop the binary variant of MPA (BMPA), a wide range of V-Shaped and S-shaped transfer functions is investigated for mapping continuous values to binary. The performance of a binary algorithm is first shown to heavily rely on the selection of an appropriate transfer function on a specific dataset. The performance of the proposed BMPA algorithm is then tested on a set of KP01 problems and compared to a number of existing algorithms. The results demonstate the merits of the BMPAs proposed in this work.