Improved GA-LNS Algorithm for Solving Vehicle Path Problems Considering Carbon Emissions

Abstract

Logistics, as a significant field for achieving energy-saving and carbon reduction goals, is recognized as a crucial direction for realizing the global “double carbon” objective, while vehicle path optimization is an effective method for promoting energy efficiency and reducing carbon emissions. In this paper, an improved genetic algorithm is proposed for optimizing logistics and distribution paths concerning the carbon emissions of fuel vehicles throughout the logistics and distribution process, and a low-carbon logistics and distribution path model is constructed based on time windows, vehicle loading, and carbon emissions. The INNC method is adopted to initialize the population, and an enhanced genetic algorithm (GA-LNS) is designed to solve the model in conjunction with a large-scale neighborhood vector search algorithm. The results indicate that the initialization of the population using the INNC method produces a higher-quality initial solution. Compared to traditional genetic algorithms and particle swarm optimization, the GA-LNS algorithm exhibits superior robustness, effectively addressing the limitations of traditional genetic algorithms that rely on initial solutions and are prone to local optima. By comparing the computational results of the low-carbon logistics distribution path model constructed in this study with those of traditional optimization objective models, it is demonstrated that this model effectively balances the trade-offs between objectives and benefits, achieving the lowest total logistics distribution cost while promoting sustainable low-carbon logistics. The research findings provide a theoretical foundation for optimizing logistics vehicle paths and formulating energy-saving and carbon reduction implementation plans in China.