Construction and Application of a Memetic Algorithm Assigning Catchment Areas to Retailers for Consumer Parcel Flow

The ongoing growth of the parcel shipping business at PostNL causes an increasing pressure on retailers that offer storage services, as storage capacity is limited. A significant share of this pressure follows from the catchment area of a retailer, the set of households for which a retailer stores certain types of parcel ow. By properly adjusting catchment areas according to the expected parcel volumes and retailer capacities, storage pressure can effectively be alleviated before turning to costly capacity expansion or acquisition of new retail locations. This thesis models the problem of determining suitable catchment areas as a Generalized Assignment Problem (GAP), minimising customer traveling distance while respecting capacity limitations at retailers. The constructed stand-alone solution method involves problem size reduction, LP-Relaxations, a Memetic Algorithm and iterated clustering, which all have been benchmarked against licenced optimisation software (Gurobi). Performance has been tested on several regions in the Netherlands, for which the results show that the method is capable of feasibly mapping catchment areas while minimising travelling distance in acceptable computation time. The method has been developed into a tool for PostNL and contributes in the decision making process of retailer catchment areas.