Optimizing blocking and starving delays in sequential zone order picking systems through time-decomposed workload balancing

Sequential zone order picking systems frequently encounter blocking delays when a tote cannot proceed to the next zone because it is occupied, and starving delays when a tote remains unassigned to a zone. General approaches to minimize delays include balancing the total workload across zones, grouping the orders into batches, or optimizing the order release sequences. However, the issue of temporal workload imbalances caused by instantaneous differences in processing time between zones has not been addressed. Since temporal workload imbalances result in delays, this study proposes the decomposition of workloads into time slots and develops a temporal workload balancing model (TBM) that incorporates batching and sequencing based on time slot decomposition. We also develop an adaptive large neighborhood search (ALNS) heuristic model to tackle large-scale practical problems of temporal workload imbalance. In simulation experiments, we compare the TBM model to alternative batching strategies in an order picking environment featuring consecutive batch windows. Our findings reveal that the TBM model yields an average reduction in makespan of 27.65% and 15.99% compared to random strategy and baseline method. We conclude that temporal workload balancing can minimize blocking and starving delays and maximize order picking productivity.