A Stochastic Optimization Approach for the Multi-Product Production Routing Problem with Multiple Uncertainties

Companies can gain significant advantages from the integrated optimization of production, inventory, and routing in a supply chain. The production routing problem (PRP), which aims to jointly optimize production, inventory, and vehicle routing, is motivated in this fashion. In actuality, the problem is typically made more challenging by the presence of several uncertainty. A two-stage stochastic mixed integer programming model is proposed under the consideration that both the production supply and the customer demand are uncertain in order to reduce the negative effects of uncertainties for PRPs. In this model, decisions regarding production and routing are made in the first stage, while decisions regarding inventory levels and lost sales are made in the second. A three-phase matheuristic algorithm is developed by integrating the mathematical programming, heuristic, and scenario reduction methods to solve this NP-hard problem. Comparing this proposed matheuristic to a commercial solver in a series of numerical experiments validates its effectiveness and efficiency. The impacts of the demand and supply uncertainty are also extensively studied. Overall, the system performance for PRP is negatively affected by the increasing demand and supply uncertainty