Failure-specific cooperative recourse strategy for simultaneous pickup and delivery problem with stochastic demands

This paper concerns the generation of a priori routes for a fleet of vehicles that pick up and deliver items with stochastic demands. A failure-specific cooperative recourse strategy is proposed to explore a risk pooling mechanism for routing in the context of simultaneous pickup and delivery with stochastic demands. By defining complete failure and semi-failure of routing, the travelling cost under our failure-specific cooperative strategy is estimated. Also, an adaptive large neighbourhood search algorithm is developed. Compared with a strategy that involves no cooperation between vehicles, our strategy performs better in terms of reducing travelling costs, and balancing fleet size and detour frequency.