An Ant Colony System Metaheuristic Algorithm for Solving a Bi-Objective Purchasing Scheduling Problem

In this work the Purchasing Scheduling Problem (PSP) is presented, based in study cases of public and private sectors. PSP models the purchasing process in both scenarios through an optimization approach. It is based on the multi-objective formulation of the knapsack problem. Therefore, PSP is defined as a based-graph problem with two objectives: maximization of satisfied demands and minimization of purchasing costs in a supplying task for inventory systems. In order to achieve these goals, the problem is defined as an integer problem, in which, feasibility of solutions is tested using a profit/cost relationship. This permits to solve PSP as a maximization of a single objective through an Ant Colony System Algorithm (ACS), an efficient solver for graphbased problems. Experimental results reveal that ACS reaches 74% of efficiency on solving instances randomly generated; obtaining purchasing plans as a result. This demonstrates the advantages of using heuristics in decision making systems such as Enterprise Resource Planning (ERP).