A hierarchical optimization technique for the strategic design of distribution networks

• We address the optimal design of the last supply chain branch, i.e., the Distribution Network (DN).
• A hierarchical procedure is presented to select the optimal DN configuration.
• Digraph modeling, mixed integer linear programming, and the analytic hierarchy process are employed.
• The optimization model is applied to a case study describing an Italian regional healthcare DN.
• The problem solution by the design method allows improving the DN flexibility and performance.

The paper addresses the optimal design of the last supply chain branch, i.e., the Distribution Network (DN), starting from manufacturers till the retailers. It considers a distributed system composed of different stages connected by material links labeled with suitable performance indices. A hierarchical procedure employing direct graph (digraph) modeling, mixed integer linear programming, and the Analytic Hierarchy Process (AHP) is presented to select the optimal DN configuration. More in detail, a first-level DN optimization problem taking into account the definition and evaluation of the distribution chain performance provides a set of Pareto optimal solutions defined by digraph modeling. A second level DN optimization using the AHP method selects, on the basis of further criteria, the DN configuration from the Pareto face alternatives. To show the method effectiveness, the optimization model is applied to a case study describing an Italian regional healthcare drug DN. The problem solution by the proposed design method allows improving the DN flexibility and performance.