Swarm Robotics

Abstract

This study analyzes and designs the Swarm intelligence (SI) that Self-organizing migrating algorithm (SOMA) represents to solve industrial practice as well as academic optimization problems, and applies them to swarm robotics. Specifically, the characteristics of SOMA are clarified, shaping the basis for the analysis of SOMA's strengths and weaknesses for the release of SOMA T3A, SOMA Pareto, and iSOMA, with outstanding performance, confirmed by well-known test suites from IEEE CEC 2013, 2015, 2017, and 2019. Besides, the dynamic path planning problem for swarm robotics is handled by the proposed algorithms considered as a prime instance. The computational and simulation results on Matlab have proven the performance of the novel algorithms as well as the correctness of the obstacle avoidance method for mobile robots and drones. Furthermore, two out of the three proposed versions achieved the tie for 3rd (the same ranking with HyDE-DF) and 5th place in the 100-Digit Challenge at CEC 2019, GECCO 2019, and SEMCCO 2019 competition, something that any other version of SOMA has yet to do. They show promising possibilities that SOMA and SI algorithms offer.