Multi-objective optimization of the order scheduling problem in mail-order pharmacy automation systems

This paper presents the multi-objective optimization problem (MOP) of minimizing collation delays and the makespan in mail-order pharmacy automation (MOPA) systems. MOPA is a high-throughput make-to-order (MTO) manufacturing system designed to handle thousands of prescription orders every day. Prescription orders are highly customized, and most of them consist of multiple medications that need to be collated before packaging and shipping. The completion time difference between the first and the last medications within the same order is defined as the order collation delay. This research mainly investigates the effects of machine flexibility and the proportion of multi-medication orders on the total collation delays. To solve this NP-hard problem, a genetic algorithm with a min-max Pareto objective function is used. The GA performance is compared to two industry heuristics: the longest processing time (LPT) and the least total workload (LTW). Experimental results indicate that a fully dedicated machine environment has 80 % more total collation delays as compared to a fully flexible machine environment and 25 % more total collation delays as compared to a multi-purpose machine environment. Experimental results also indicate that the GA can achieve the optimal makespan in most cases, while minimizing the total collation delays by 96 % when compared to LPT and LTW.