Abstract
The latency location routing problem integrates the facility location problemand the multi-depot cumulative capacitated vehicle routing problem. Thisproblem involves making simultaneous decisions about depot locations andvehicle routes to serve customers while aiming to minimize the sum of waiting(arriving) times for all customers. To address this computationally challengingproblem, we propose a reinforcement learning guided hybrid evolutionaryalgorithm following the framework of the memetic algorithm. The proposedalgorithm relies on a diversity-enhanced multi-parent edge assembly crossoverto build promising offspring and a reinforcement learning guided variableneighborhood descent to determine the exploration order of multipleneighborhoods. Additionally, strategic oscillation is used to achieve abalanced exploration of both feasible and infeasible solutions. Thecompetitiveness of the algorithm against state-of-the-art methods isdemonstrated by experimental results on the three sets of 76 popular instances,including 51 improved best solutions (new upper bounds) for the 59 instanceswith unknown optima and equal best results for the remaining instances. We alsoconduct additional experiments to shed light on the key components of thealgorithm.