Clustering and routing in waste management : a two-stage optimisation approach

This paper proposes a two-stage model to tackle a problem arising in Waste Management. The decision-maker (a regional authority) is interested in locating sorting facilities in a regional area and defining the corresponding capacities. The decision-maker is aware that waste will be collected and brought to the installed facilities by independent private companies. Therefore, the authority wants to foresee the behaviour of these companies in order to avoid shortsighted decisions. In the first stage, the regional authority divides the clients into clusters, further assigning facilities to these clusters. In the second stage, an effective route is defined to serve client pickup demand. The main idea behind the model is that the authority aims to find the best location-allocation solution by clustering clients and assigning facilities to these clusters without generating overlaps. In doing so, the authority tries to (i) assign the demand of clients to the facilities by considering a safety stock within their capacities to avoid shortages during the operational phase, (ii) minimise Greenhouse Gases emissions, (iii) be as compliant as possible with the solution found by the second stage problem, the latter aiming at optimising vehicle tour lengths. After properly modelling the problem, we propose a matheuristic solution algorithm and conduct extensive computational analysis on a real-case scenario of an Italian region. Validation of the approach is achieved with promising results.