Metaheuristic optimization algorithms for single mobile robot scheduling

Abstract

U ovom radu se analizira problem terminiranja mobilnog robota (MR) u cilju pronalaženja optimalnog načina opsluživanja mašina alatki u inteligentnom tehnološkom sistemu. Da bi se rešio ovaj NP-hard optimizacioni problem, koriste se različiti biološki inspirisani metaheuristički algoritmi optimizacije, poput algoritama inspirisanog inteligencijom roja čestica - Particle Swarm Optimization (PSO), algoritma inspirisanog inteligencijom jata kitova - Whale optimization algorithm (WOA) i modifikovanog WOA (mWOA). Najbolje performanse pokazao je algoritam mWOA, koji je biološki inspirisan algoritam zasnovan na principu lova jata grbavih kitova. Inteligentni tehnološki sistem korišćen za eksperimentalnu verifikaciju predloženog algoritma se sastoji od osam mašina alatki na kojima se vrši obrada tri dela, koji imaju pet tipova fleksibilnosti, dok se terminirani unutrašnji transport materijala u tehnološkom sistemu vrši jednim mobilnim robotom. Razvijeni algoritam je implementiran u MATLAB softverskom paketu

In this paper, the authors analyze single mobile robot scheduling problem in order to find an optimal way to transport parts in intelligent manufacturing system. Because of the combinatorial complexity, this problem is considered to be NP-hard and the authors propose three different metaheuristic algorithms (PSO, WOA and mWOA) to solve it. Manufacturing system consists of eight machine tools, three parts and single mobile robot used for material transport tasks. Optimal scheduling plans are obtained by single objective optimization procedure, using four fitness functions. The experimental results show that mWOA outperforms all other algorithms. All the algorithms are implemented in MATLAB software package