An application to stochastic vehicle-routing problem in a waste collection

In this research, we consider a dynamic collection plan with stochastic demands. Each customer has a known position and a random demand. Each vehicle has the same capacity. This problem differs from the well-known stochastic vehicle-routing problem with that re-pickup occurs if the capacity constraint of a vehicle is exceeded. This problem arises in several situations such as waste collection and less-than truckload motors. Many stochastic vehicle-routing problems deal with issues related to customer delivery. In this case, a stock-out occurs owing to the variation of demand. Several studies deal with a stock-out risk as a constraint. The studies are added to the objective function as a penalty. However, in this research, it is necessary to add the excess distance to the objective function by re-pickup. The aim of this research is to model the dynamic collection plan by considering the variation of demand and develop a search algorithm. In this research, the vehicle-routing problem that assigns n customers to K vehicles can be formulated as a combinatorial optimization problem. The proposed search algorithm is applied to an actual trash collection problem.