Genetic algorithms with greedy strategy for green batch scheduling on non-identical parallel machines

Large scale batch scheduling problems with complex constraints are difficult and time-consuming to solve. Therefore, this paper addresses the green batch scheduling problem on non-identical parallel machines with time-of-use electricity prices. The objective of the problem is to minimise total electricity costs (TEC) in production. Two kinds of algorithms—single-population genetic algorithms (SPGA) and multi-population genetic algorithm (MPGA)—are proposed to solve the problem. In the algorithms, the products are allocated into batches and are then allocated to machines randomly. A greedy strategy is designed to arrange the production sequence and the starting time of the batches. Furthermore, a self-adaptive parameter adjustment strategy is proposed to enhance the adaptability of the algorithm. Computational experiments with CPLEX solver have been conducted to evaluate the performance of the algorithms. On small instances, both SPGA and MPGA can achieve approximate results compared with those obtained by CPLEX, and can also achieve smaller TEC on large instances with less computing time. In addition, the proposed MPGA implemented by parallel computing outperforms SPGA in getting better results with nearly the same computing time.