Genetic algorithms for scheduling optimization on a multi-product batch processing machine

This paper describes loading polices for the scheduling problem on a multi product batching processing machine(BPM) which can process a batch of jobs simultaneously with a known and fixed number of jobs and their ready time. The different jobs are dispatched and sequenced in order to minimize the makespan and maximize the utilization of the servers. As an extension to the basic model of previous work (Fanti et al, 1997) for BPM scheduling, we build up the model to schedule n jobs on m identical servers in which the optimization procedure results in a complex NP-hard combinatorial problem. Genetic Algorithms(GAs) are applied to solve this scheduling problem where we apply the features of elitist strategy GAs to develop a group of MATLAB functions for solving the BPM scheduling problem. This result is the optimal solution. This experiment demonstrates that GAs can provide a robust search procedure in the optimization of scheduling problem which has high dimensionality, multi-modality, discontinuity and noise (DeJong, 1975).