Genetic algorithms for parallel machine scheduling with setup times

This paper discusses the problem of scheduling a set of jobs on parallel machines considering sequence-dependent setup times. The processing sequence of jobs assigned to each machine is determined to minimize the makespan. Since it is a NP-hard problem, even for instances with small sizes, the computational effort may be prohibitively large. Then genetic algorithms with two types of assignment scheme are presented to obtain near optimal solutions. The popular scheme of random assignment is used in the genetic algorithm (GA-1). Unlike GA-1, the scheme of greedy assignment is employed in the genetic algorithm (GA-2). The computational experiments are carried out to compare the performance of GA-1 with that of GA-2. Computational results show that GA-2 has better performance than GA-1.