Computational Experiments on Workflow Balancing of Parallel Machine Scheduling with Precedence Constraints and Sequence Independent Setup Time

The workflow balancing of parallel machines scheduling (PMS) with precedence constraints and sequence independent setup time is considered for study. The setup time consideration produces alternate schedule along with lesser relative percentage of imbalance (RPI) value for PMS problem is demonstrated with an example. The lesser RPI indicates better workflow balancing among machines. The computational experiments are conducted on large instances of randomly generated PMS problems with precedent constraints and setup time. The various combinations of heuristics are used to solve the problems. The results show that genetic algorithm (GA) performs well against the other heuristics with lesser RPI values