Applying a modified adaptive large neighbourhood search for truck scheduling and pile assignment in a two-stage sorting system

In this study, we tackle the complexities of a two-stage semi-automatic sorting system, considering the diverse distribution requirements of parcels and the constraints imposed by sorting equipment. Our objective is to integrate two decision points - the inbound truck schedule and the parcel sorting plan - to minimise overall operational costs. We first formulate the problem using a mixed-integer linear programming model and then propose a mixed-coded modified adaptive large neighbourhood search (MCMALNS) algorithm to enhance performance. In our computational study, the proposed approach demonstrated the ability to quickly obtain high-quality solutions compared to other algorithms. Furthermore, a full factorial experiment was conducted to analyse cost variations across 36 scenarios. Factors including loading, deadline, arrival pattern, pile/commodity ratio, and algorithm were all identified as significant and exhibited considerable influence on the outcomes. The insights derived from this analysis provide valuable guidance for management personnel in decision-making. [Received: 30 January 2023; Accepted: 25 August 2023]