Heuristic approaches for last-mile delivery optimization

Last-mile delivery represents the final and most costly segment of the supply chain, accounting for a significant portion of total logistics expenditure. As e-commerce continues to expand globally, the pressure to optimize lastmile operations has intensified. Exact optimization methods, while ensuring optimal solutions, are computationally intractable for large-scale real-world instances. Heuristic approaches offer a practical alternative by producing good quality solutions within acceptable timeframes. This article reviews the principal heuristic methodologies applied to last-mile delivery problems, covering classical construction heuristics, metaheuristics, and hybrid approaches. The review highlights significant algorithmic contributions, discusses their strengths and limitations, and outlines new direction including the integration of machine learning with heuristic. Findings suggest that heuristics remain essential tools for practitioners and researchers seeking efficient, scalable delivery route planning.