A Two-Phase Heuristic for the Collection of Waste Animal Tissue in a Colombian Rendering Company

Abstract

This work addresses the planning of the collection of waste animal tissue in a Colombian rendering company. Over a week, the rendering company visits more than 800 slaughterhouses, butchers, and supermarkets in the Aburra’s Valley, the metropolitan area of Medellin (Colombia) to supply their plant (located in the outskirts of the city) with raw material that are transformed into value-added products. The underlying vehicle routing problem have several distinguishing features: periodicity, consistency, clustered customers and heterogeneous fleet. To solve this rich VRP we present a two-phase heuristic. The first phase of the heuristic groups the collection points using a capacitated concentrator location problem (CCLP). Then, in the second phase a mixed integer program schedules the visits of the collection points in each cluster to balance the number of visits performed daily based on the capacities of the available vehicles. These two phases aim at getting consistent and evenly spread visits during the week. Preliminary results with the data of the current operation reveal a savings potential of 5 out of 15 vehicles, and a better spread of the visits over the planning horizon.