A multi-level capacitated arc routing problem with intermediate facilities in waste collection

This study investigates a multi-level capacitated arc routing problem for waste collection with two levels of intermediate facilities: huts and transfer stations. The problem aims to implement integrated optimisation of the intermediate facilities and find optimal routes for a fleet consisting of manually operated vehicles and vehicles with compressors. We first build an integer linear programming model to address this problem. Then, we propose an extended adaptive large neighbourhood search (E-ALNS) approach, including a tailored SHAKE procedure inspired by variable neighbourhood search to effectively solve this problem. Computational experiments were conducted on 264 instances generated using existing benchmarks. Compared to commercial optimisation software and other algorithms, our extended algorithm exhibits superior performance for this problem. The results also demonstrate that the integrated optimisation of intermediate facilities significantly decreases the total cost compared to separate optimisations.