Optimization of the structure of the parcel distribution system at a selected logistics company

The article presents a description, formulation and method of solving a certain problem of optimization of the structure of the distribution system in a selected logistics company in Poland. It was assumed that shipments are generated in the areas of address codes (more precisely: we know the volume of shipments between each pair of address codes in a time profile (day/week/month/year)). The problem of system structure optimization is decompressed into two subproblems: (1) location and dimensioning the company's branches by optimizing the number, location, bandwidth of branches and allocating code points to be serviced by the branches. The criterion for optimization is the total cost of establishing and maintaining branches and the costs of transporting shipments. From code points to the branch and vice versa; (2) location and dimensioning of the sorting plant and determination of the transport plan. Problem (1) solved using a modified, randomised clustering method k-means on the current data on the structure and volume of shipments from a certain courier company. Results of research on real data for the problem (1) are presented.