A Periodic Routing Problem with Stochastic Demands

Increasingly demanding requirements in logistics and transportation plannings necessitate advanced models. This paper introduces a formal model for the distribution of goods to customers with stochastic demands aiming at timely and entire order fulfillment and utilizing constrained resources efficiently. For such (multi-)periodic routing problems with stochastic demands the sequence of customer visits on the basis of day tours for a one week planning period needs to be determined in advance. The problem type under consideration is constrained by truck capacity, driving time and customer preferences for the frequency of deliveries. In a computational study two metaheuristics are applied to 18 problem instances based on real world data and to 36 synthetic instances with up to 50 customers varying capacities, clustering and demand properties systematically.