Optimization Methods for Vehicle Routing in Transportation Planning and Logistics: A Structured Review and Algorithm-Selection Framework

Vehicle routing lies at the core of transportation planning and logistics, and a large body of Operations Research (OR) methods has been developed to solve it. This paper presents a structured narrative review of optimization methods for the capacitated vehicle routing problem (CVRP) and its principal extensions. The review covers mixed-integer programming formulations, exact branch-and-cut and branch-and-price algorithms, classical construction heuristics, modern metaheuristics (tabu search, simulated annealing, genetic algorithms, ant colony optimization, particle swarm optimization, adaptive large neighborhood search and hybrid genetic search), matheuristics and learning-based approaches, using a compact mixed-integer linear programming (MILP) model as a common reference formulation. We propose a classification framework that organizes methods by solution guarantee, search paradigm and suitability for different problem scales, and derive practical guidelines for algorithm selection based on instance size, constraint structure, uncertainty and available computing time. We also summarize standards for fair computational comparison, as inconsistent benchmarking is a recurrent weakness of the applied metaheuristic literature. No single method dominates: exact methods remain indispensable for small or well-structured instances and for computing bounds, whereas well-engineered hybrid metaheuristics are preferred at scale, and several population-based algorithms popular in applied studies are generally less competitive than neighborhood-search-based methods on standard benchmarks. Open research directions include integrating machine learning with optimization, dynamic and stochastic routing, and reproducible benchmarking.

کلیدواژه ها:

operations research ، vehicle routing problem ، Mixed-Integer Programming ، Metaheuristics ، transportation planning ، Logistics