Study On Improved Ant Colony Algorithm And Its Application In The Vehicle Scheduling

ABSTRACTWith the development of market economy, it has been paid more attention that logistics, taken as the third profit resource, has obviously affected the economy activity. As the most important competitive field, it will play an important role in the future market competition. The vehicle scheduling optimization and system of logistics delivery is one of important approaches that implement fast, accurate and low-cost logistics delivery. And they are indispensable parts of modern logistics system.In current logistics system, manual or computer-aided approaches are mostly adopted. On one hand, it will cost much time, on the other hand it is hardly to synthesize multi-objective and multi-constraint scheduling demands to perform scientific quantitative analysis and optimization disposal. So it is desired for theory and practice to present effective and adaptable intelligent optimization for general vehicle scheduling problems, research on vehicle scheduling requirement in logistics delivery, establish the mathematical model of vehicle scheduling in complicated environment with multi-objective and multi-constraint. The main contributions of this thesis include:1．The theory, model, tool and simulation of basic ant colony algorithm are discussed in details by analyzing and comparing intelligent routing optimizations.2．Two novel intelligent optimizations, named Dynamic Little Window Ant Colony Algorithm based on Pattern Learning (DLWACAPL) and DLWACAPL integrated Genetic Algorithm (HACAGA), are proposed to avoid the deficiency of basic Ant Colony Algorithm which often costs long time and get into the local optimization easily. And the validations and adaptation of the two intelligent optimizations have been performed with benchmark.3．The mathematical models and approaches of two typical vehicle scheduling problems on Capacitated Vehicle Routing Problems (CVRP) and Vehicle Routing Problems with Time Windows (VRPTW) are investigated. Furthermore, the multi-objective optimization strategies on vehicle scheduling including multi-objective framework, mathematics model and multi-objective comprehensive optimization are researched.4．Vehicle scheduling system based on the improved ant colony algorithms (DLWACAPL and HACAGA) has been developed, including the design of main function models, the configuration of software and hardware as well as coding and realization of the key algorithms.5．A study of how these meta-heuristics perform is carried out on the engineering benchmark. And the experimental results have indicated the validation and adaptation of the improved ant colony algorithm proposed in this thesis.Lan Shihai (Mechanical and Electrical Engineering)Supervised by...