Stochastic dynamic vehicle routing in the light of prescriptive analytics: A review

Stochastic dynamic vehicle routing problems have become an essential part of logistics and mobility services. In such problems, a sequence of vehicle routing decisions has to be made in reaction and anticipation of newly revealed stochastic information. To this end, a variety of computational operations research methods has emerged in the literature, increasingly integrating potential future information in their decision making. The integration of information models into decision models via computational methods is known as prescriptive analytics, the most recent advance of business analytics. In this paper, we explore the existing work and future potential of prescriptive analytics for stochastic dynamic vehicle routing. We identify the characteristics of decision models and information models unique in stochastic dynamic vehicle routing and analyze how different methodology meets the characteristics' requirements. We use the insights to derive recommendations about promising methodology when approaching specific stochastic dynamic vehicle routing problems.