Dynamic integrated optimization of batching and routing in narrow-aisle order picking systems with congestion consideration

Abstract

The e-commerce industry has driven logistics centers to process massive and time sensitive orders more efficiently and effectively. However, the high land and construction costs have forced the warehouse to adopt a narrow channel layout to increase its capacity, which in turn has caused congestion and further reduced its operational efficiency. Therefore, how to carry out optimization of order batching and picking paths in narrow channels to avoid channel congestion and pursue a lowest total cost has become the problem that this article aims to solve. This article develops a Markov decision process model to address the online order multi-period picking planning optimization challenge, and proposes the Genetic Algorithm-Ratliff Rosenthal-Time Weighted Similarity (GA-RR-MTWS) algorithm which innovatively integrates congestion considerations into the optimization process. A myopic cost function approximation strategy is introduced aiming at minimizing the total cost of a whole working day. Comparative experimental analysis demonstrates the modified cost function GA-RR-MTWS's superior performance in reducing total picking cost and congestion, particularly in complex, multi-aisle environments with multiple pickers. The method's ability to manage congestion and optimize routing significantly improves overall warehouse efficiency.