Otimização da cobertura de dispensação de medicamentos com minimização de sobras: um sistema baseado em algoritmos genéticos

Abstract

High-cost medication centers face a critical logistical dilemma: ensuring full coverage of scheduled patients while minimizing waste due to product expiration. In this context, this work proposes a computational system based on a Genetic Algorithm (GA) to support inventory planning, using the Juarez Barbosa State Center for High-Cost Medication as reference. Due to the unavailability of real data from the center, all executions were performed with simulated datasets designed to accurately reflect the operational dynamics of such environments — making the model applicable to any center with scheduling and shelf-life management. The problem was modeled as a variant of the Lot-Sizing and Scheduling Problem with Shelf-Life Constraints, where mandatory coverage of scheduled demand is a critical constraint, and the goal is to minimize surplus stock at risk of expiring before the next acquisition. The GA was integrated with a Monte Carlo simulation model accounting for patient adherence uncertainty. Experiments showed that the model converges to 100% coverage when the initial stock allows full demand fulfillment, remaining stable even under changes in the objective function weights. In more challenging scenarios with limited stock, average coverage ranged between 96.0% and 97.1%, with up to 28% reduction in surpluses when efficiency was prioritized. It is concluded that the proposed system is robust, scalable, and suitable as a predictive tool for sustainable logistics optimization in high-cost medication centers.