Stochastic Vehicle Routing Problem: classification and review

This paper reviews recent research on the Stochastic Vehicle Routing Problem (SVRP), where key parameters such as customer demand, requests, travel time, and service time are uncertain. The literature is organized according to the type of stochastic parameter considered and the modeling approaches used. Four major modeling frameworks are discussed: two-stage stochastic programming. chance-constrained programming, robust optimization, and fuzzy modeling. To illustrate how these approaches are applied, concise formulations for each framework are presented. The review focuses on studies published after 2018 and analyses 78 papers in total. For SVRPs with stochastic demand, studies are further classified by their recourse strategies: no recourse, classical recourse, and optimal recourse. The review highlights research trends, showing that demand uncertainty is the most frequently studied, followed by travel and service time uncertainty. Increasingly, studies combine multiple uncertainties and employ advanced solution methods such as heuristics, decomposition methods, simulation, and machine learning techniques.