Acceptance and dispatching policies for a distribution problem

A dynamic and stochastic distribution problem with a number of terminals and a fleet of vehicles is analyzed. Customers request the transportation of batches of loads between different origins and destinations. A request can be accepted or rejected; if the request is accepted, a reward is received. Holding costs for vehicles and loads at terminals as well as transportation costs are included in the model. The objective is to determine a policy for accepting transportation requests and for dispatching vehicles that maximizes the expected value (rewards minus costs) of operating the distribution system. A Markov decision process model is developed, optimal policies are characterized, and algorithms that exploit the structure of the problem are developed.