Less Expensive Formulation for a Realistic Routing-Scheduling-Loading Problem (RoSLoP)

摘要

In this paper the Routing-Scheduling-Loading Problem (RoSLoP) is approached.This is a rich bin packing (BPP) and vehicle routing (VRP) problem formulated to satisfy thetransportation requirements of a bottling company located in Mexico. The initial formulation ofthe problem uses e integer variables and 30 constraints making difficult to find the exact solu-tion even for small instances. In this work it is proposed a transformation function that reducesthe size of the problem formulation which allows obtaining the optimal solution of small in-stances using an exact algorithm. Experimental results of the performance evaluation of an ap-proximated solution method, with regard to the optimal solution, are showed. It is important toemphasize, that this is the first time that this kind of evaluation is carried out for RoSLoP. Inthe experiments a set of 12 test instances were selected from the company database. Theexperimental evidence shows that the transformation function reduces 97% the number of cus-tomers orders. The percentage quality error for the traveled distance was 0% and for the vehi-cles used was 6.19%. Now these results can be used to evaluate the performance of any newapproximation solution method of RoSLoP.