OTIMIZAÇÃO DA ALOCAÇÃO DE POSTOS DE RESFRIAMENTO DE LEITE POR MEIO DE HEURÍSTICA BASEADA EM MCKP E VALIDAÇÃO MILP

Abstract

This work presents a methodology for optimizing the allocation of milk cooling stations in the state of Goiás, considering logistical, operational, and geographical aspects. Productive dispersion, associated with inadequate station locations and the length of collection routes, increases logistical costs, reduces operational efficiency, and can compromise product quality. Given this scenario, the use of mathematical models capable of supporting strategic decisions related to the installation and sizing of these structures becomes essential. The objective of this study is to develop a computational model, implemented in Python, capable of defining the optimal location of cooling stations, minimizing the total logistical cost, considering transportation costs, installed capacities, and fixed implementation costs. For this, a hybrid approach based on the Multiple-Choice Knapsack Problem (MCKP) heuristic computational method, validated by a Mixed-Integer Linear Programming (MILP) model, is used. The methodology incorporates routing techniques using the Clarke-Wright and 2-opt methods, as well as the use of real distances obtained through the Open Source Routing Machine (OSRM). The results demonstrate that the heuristic model performed near optimally, with a difference of less than 2% compared to MILP, but with significantly less computational time. Furthermore, the method proved to be scalable, maintaining efficiency even in scenarios with a high number of producers. Thus, the proposed approach is an effective tool to support strategic decisions, optimize logistics costs, and contribute to the planning of the milk production chain.