A hybrid genetic algorithm for parallel machine scheduling with setup times

Abstract This paper addresses the unrelated parallel machine scheduling problem with sequence and dependent setup times eligibility constraints. The objective is to minimize maximum completion time (makespan). Instances of more than 500 jobs 50 machines are not uncommon in industry. Such large instances become increasingly challenging provide high-quality solutions within limited amount computational time, but so far, have been adequately addressed recent literature. A hybrid genetic algorithm developed, which lean sense that equipped a minimal number parameters operators, enhanced an effective local search operator, specifically targeted solve instances. For evaluation purposes new set larger problems generated, consisting up 800 60 machines. An extensive comparative study shows proposed method performs significantly better compared other state-of-the-art algorithms, especially for Also, it demonstrated calibration crucial practice should be at narrower representative