Simulation-Based Optimization Approaches for Dealing with Dual-Command Crane Scheduling Problem in Unit-Load Double-Deep AS/RS Considering Energy Consumption

Abstract

Automated storage/retrieval systems (AS/RS) have been increasingly used to support operations in manufacturing firms, warehouses, and distribution centers. Usually, AS/RSs are expensive. To achieve a good return on investment (ROI), an AS/RS must operate optimally. This research focuses on solving the crane scheduling problem, which has a great and immediate impact on the performance of an AS/RS. To optimize the design and operations of an AS/RS, many past studies have applied the simulation approach. However, the simulation and optimization have been often loosely coupled, resulting in a rigorous and labor-intensive optimization procedure. Using population- and evolution-based metaheuristics to deal with the crane scheduling problem of an AS/RS is one of the research trends. However, the whale optimization algorithm (WOA) and its variants have not been used for this purpose. To address the said gaps, this research first proposes a framework for coupling the simulation and optimization closely, in which various heuristics/metaheuristics, including first-come first-serve (FCFS), RANDOM, WOA, genetic algorithms (GAs), particle swarm optimization (PSO), and especially an improved WOA (IWOA), together with dynamic programming (DP), have been used as alternative sequencing methods. Based on this framework, different simulation-based optimization approaches have been developed for solving the dual-command crane scheduling problem in a unit-load double-deep AS/RS. The experimental results show that IWOA+DP outperforms the others in terms of energy consumption.