SWARM INTELLIGENCE: APPLICATION OF THE ANT COLONY OPTIMIZATION ALGORITHM TO LOGISTICS‐ORIENTED VEHICLE ROUTING PROBLEMS
研究比较了蚁群优化算法与Clark-Wright节约算法在物流车辆路径问题上的表现，发现蚁群算法在20个需求点以内的问题中能找到接近最优的解，且优于节约算法，建议在更大规模问题中测试。
This research evaluates a set of logistics‐oriented vehicle routing problems (VRP) taken from the logistics and supply chain literature under the widely used Clark‐Wright Savings algorithm and the newer metaheuristic method employing a type of swarm intelligence called Ant Colony Optimization (ACO). ACO simulates the decision‐making processes of colonies of ants as they forage for food and is related to other artificial intelligence techniques such as Tabu Search, Simulated Annealing and Genetic Algorithms. Experimentation shows that ACO is successful in finding solutions near the best‐known solutions for problems with up to 20 demand locations. In addition, testing for the affect of spatial patterns suggested by the logistics literature for facility locations appears to make a difference in the quality of the solutions for the two algorithms. Finally, ACO is shown to be superior to the savings algorithm found in software packages and as a result should be tested on even larger, more complex logistics‐oriented vehicle routing problems, representative of those encountered in larger industrial and retail settings.