Multi-Swarm-based Parallel Spider Monkey Optimization Algorithm

Abstract

Particle swarm optimization (PSO) algorithms face performance challenges, which lean on improving solutions quality, speed-up, dealing with large-scale problems, and exploitation of computational resources. Parallelism represents a suitable paradigm for overcoming the PSO challenges. Spider monkey optimization (SMO) algorithm is a recent PSO algorithm. SMO is based on the principle of dividing the swarm into subgroups, which may decrease its speedup. In this paper, a multi-swarm-based parallel spider monkey optimization (PSMO- MS) is proposed for dealing with large-scale problems based on the multi-swarm mechanism. PSMO-MS is implemented using a synchronous master/slave parallel model. The performance of the proposed PSMO-MS is tested on two 2-dimensional problems (Dekkers and Aarts problem and Camel Back-6 Six Hump problem). The results show that PSMO-MS outperforms SMO in terms of execution time and produces comparable and better solution quality with a large-scale problem, as well as a high solutions’ density.