Rural last-mile distribution using a hybrid heuristic-multi-criteria decision-making framework

Autor(es)

Rodrigues, Gustavo de Abreu ; Junqueira, Leonardo ; Vieira, Jose Geraldo Vidal ; da Silva Junior, Orivalde Soares ; da Cunha, Claudio Barbieri

• Tipo de documento: Artigo Científico
• Fonte: EXPERT SYSTEMS WITH APPLICATIONS; v. 289, p. 19-pg., 2025-09-15.

Resumo

In this paper, we address the Rural Last-Mile Distribution Planning Problem (RLMDPP), a real-world variant of Rich Vehicle Routing Problems (RVRP) that arises in the context of delivering supplies to dispersed customers located on farms in rural areas. The objective is to develop daily dispatch plans that integrate inventory allocation and vehicle routing under multiple operational constraints. To tackle this problem, we propose a novel hybrid solution framework that combines sequential heuristics with a Multi-Criteria Decision-Making (MCDM) approach based on Multi-Attribute Value Theory (MAVT). The framework generates multiple dispatch plan alternatives and evaluates them using the SMARTER technique, which incorporates decision makers' preferences into the ranking process. We apply the proposed method to a real-world case from a Brazilian company operating in the rural feed supply sector. Computational experiments show that the framework significantly outperforms manually generated plans by experienced logistics professionals in terms of cost-efficiency, service level, and tender acceptance chance, producing optimized plans in short running times, thus highlighting the potential of integrating MCDM techniques into vehicle dispatching and routing to support enhanced decision-making in complex distribution environments. (AU)

• Processo FAPESP: 22/05803-3 - Problemas de corte, empacotamento, dimensionamento de lotes, programação da produção, roteamento e localização e suas integrações em contextos industriais e logísticos
• Beneficiário: Reinaldo Morabito Neto
• Modalidade de apoio: Auxílio à Pesquisa - Temático
• Processo FAPESP: 22/03182-1 - Modelos matemáticos e métodos de solução para problemas de otimização combinatória de grande porte
• Beneficiário: Leonardo Junqueira
• Modalidade de apoio: Auxílio à Pesquisa - Projeto Inicial