You are here

Real – time Management of Ex van Deliveries

In this dissertation a basic case of the Vehicle Routing Problem (VRP) is studied, in which a single vehicle starts from its depot and serves customers in a predefined sequence. The objective is to serve all customers and minimize travel distance (cost). This problem is of significant practical interest; indicative applications include Ex-Van sales and Material Handling systems. Several cases of this problem, of increasing complexity, are posed, analyzed and solved. These cases are:

  • Multiple product delivery with deterministic customer demand. Two sub-cases are studied: a) the compartmentalized load and b) the unified load case. The mathematical models, as well as new efficient algorithms that solve these problems to optimality have been developed and analyzed.
  • Multiple product delivery with stochastic customer demand. Both sub-cases mentioned above are studied. For both cases we present the characteristics of the respective problems, novel methods to determine the minimum expected cost, and the theoretical results that permit one to determine the optimal decision after serving each customer. Both cases have been addressed using dynamic programming, and for both it has been proven that there exists an appropriate threshold function for each customer, which can be used to determine the optimal decision. Extensive analysis of the proposed algorithms has been conducted.
  • Pickup and delivery (of product) with stochastic customer demands. In this case the vehicle not only delivers products to customers but it also picks up returned items from each customer (e.g. damaged goods, or empty packaging). The characteristics of the problem have been presented, together with a novel method to determine the minimum expected cost, and the optimal decision after serving each customer. The proposed method has also been analyzed extensively.

This work may support a decision support framework, which can be utilized in fixed routing operations for a wide variety of cases and applications (deterministic or stochastic demand, single or multiple products, delivery or pickup & delivery): Thus, ad-hoc sub-optimal decisions can be eliminated, minimizing total operating costs, and increasing the overall productivity and customer service of the distribution fleet.