MILP for a Variant of Pickup & Delivery Problem for both Passengers and Goods Transportation

Pickup and delivery problems have been declined in many variants. In these optimization problems, a fleet of vehicles must satisfy customers' transport demand under specific constraints. This paper studies the feasibility of using a linear model including constraints that are not only related to the goods transport, but also combined with passengers transport in a more flexible way. We explore many characteristics simultaneously like transfers capabilities (with and without storage), time windows on sites, on transfer points and on vehicle depots, heterogeneous capacities and multiple visits. We use exact methods on some instances we created, using commercial solver IBM CPLEX. Results obtained show potential savings on two instances but also highlight the complexity and limits in terms of scalability.