A novel covering approach to positioning ELV collection points

This paper presents modelling approach that could be used to establish one important part of reverse logistics networks for end-of-life vehicles (ELVs) by defining optimum locations for collection points. Determining location of collection points in ELV network usually includes existence of one or more allowable service distances between last owners and collection facilities, as well necessity of serving each of them. Namely, each vehicle owner can be considered as potential demand point in accordance to some predefined standards. Fact that the number of residents in urban areas is usually very large makes the location problem difficult to solve. To avoid such a large number of demand points, it is a common practice to aggregate them. This would mean to represent all vehicle owners from certain area as one demand location, which in turn may cause errors while configuring collection network. Also, existence of more than one allowable service distances creates additional service zones around collection facilities, which increases problem complexity. Therefore, to minimise aggregation errors, this research proposes approach to partition service zones into sub zones, i.e., subsets of demand locations. Sub zones are covered by one or more collection facilities and demand of each sub zone is proportional to the covered zone area. All of these modifications are incorporated into the novel formulation of maximal covering location problem. The proposed modelling approach is illustrated on Belgrade city area and obtained results confirmed that it can be used very efficiently to position ELV's collection facilities.