Seleção Inteligente de Heurísticas para o Problema do Empacotamento com Bins Variáveis e Conflitos: Uma Abordagem Baseada em ´ Arvore de Decisão

Abstract

In this article, we analyze the Variable-Sized Bin Packing Problem with Conflicts (VSBPPC). This problem involves allocating items of specific si zes into bins with varying capacities and costs, while respecting conflict cons traints between item pairs, with the goal of minimizing the total cost of the bins used. Given its practical relevance in contexts such as resource alloca tion, production planning, and logistics distribution, and motivated by the high computational complexity of existing approaches, we present enhancements to the Large Neighborhood Search Algorithm (LNSA), including variations in the constructive heuristic, solution reconstruction strategies, and the integration of machine learning models to guide the algorithm’s decisions. The proposed ap proach was compared to the LNSA described in the literature, demonstrating an average performance improvement of 0.085% while reducing execution time to approximately one-third of that required by the original algorithm.