Font Size: a A A

The Research On Incident Vehicle Routing Problem

Posted on:2016-01-07Degree:DoctorType:Dissertation
Country:ChinaCandidate:Y L TangFull Text:PDF
GTID:1108330482955260Subject:Control theory and control engineering
Abstract/Summary:PDF Full Text Request
The high-speed development of social economy puts forward higher requirements on modern logistics industry. There is a situation in real life, customers need different parts, and the parts are part of the finished goods, due to the differences in the nature, characteristics and uses, but there exists a relationship between them. And customers will entrust to a logisitic company to finish the supply business in order to ensure the demand, the goods are compatible, and this kind of delivery way is conducive to the latter operation and business, based on these situation, we propose incident vehicle routing problem (IVRP). In this paper, several kinds of IVRP with extended features mathematical model are established, correspondingly, several algorithms are constructed to solve these problems.The main work is described as follows:(1) Chaos genetic algorithm for single-depot hmogenic for IVRP is proposed. Give chaos disturbance for bad gene, reduce the search space and improve the optimization speed. Analyze the formation mechanism and change rule for the solution, the simulation results prove the effectiveness of the proposed algorithm.(2) Hybrid tabu search algorithm for single-depot multi-depot IVRP. is proposed. Combine the advantages of the tabu search and genetic algorithm, core path, adaptive crossover and chaotic mutation are introduced, the influence of road condition is considered, apply the proposed algorithm to solve 20 customer scale, give comparison of convergence time, evolution and global search probability, verify the feasibility of the proposed algorithm.(3) Hybrid ant colony optimization for multi-depot heterogeneous IVRP with many types of time windows is proposed. A mathematical model is established. Construct the initial solution by clustering algorithm and saving algorithm, improve the speed of solving; change heuristic factor and expectation heuristic factor adaptively, improve the convergence speed; genetic operator is introduced, and change the crossover probability and mutation probability adaptively, improve the global search ability; extracting core path is benefit for latter optimization; improve the local search ability by 3-opt and depot exchange. The simulation results for 40 customer scale proved the superiority of the proposed algorithm.(4) Adaptive ant colony optimization for IVRP under time-varying network environment is propsed. Considering these factors in the real life, such as road conditions affect the transportation cost, fuel consumption rate is relative to the load, dealing with crossing multi-period problem, vehicle routing problem in time-varying networks environment mathematical model is established. Constructing initial solution through cluster algorithm and saving algorithm, changing information heuristic factor and pheromone expectation heuristic factor, combining the fuel consumption rate to update pheromone adaptively, ensuring the convergence speed according to the above methods. At last,3-opt strategy was used to improve local search ability. The simulation result shows that the proposed algorithm is feasible.(5) A hybrid ant colony collaborative algorithm for domain IVRP is proposed. We present two kinds of domain-oriented application problems, namely school bus and rural public transportation collabrative IVRP, based on genetic algorithm and ant colony optimization, to construct a kind of hybrif ant colony collaborative algorithm, chaotic search can overcome the defect of generating large amounts of unfeasible solution, accelerate chromosomal convergence to the optimal solution, a smooth mechanism can improve the ability of exploration data processing by increasing the the probability of choosing low intensity pheromone of element, the simulation result of two kinds of model shows that the searching capability of the proposed algorithm is improved.(6) A parallel hybrid ant colony algorithm for IVRP is proposed. On the basis of in-depth analysis, using the proposed algorithm to solve the established mathematical model. The experimental results show that the proposed is better than the ant colony optimization algorithm in dealing with large-scale model.At last, summing up the thesis and looking forward to the future, achievements are concluded, and the problems need to discuss deeply in the near future are pointed out.
Keywords/Search Tags:Incident Vehicle Routing Problem, time wondows, time-varying network, genetic algorithm, ant colony optimization
PDF Full Text Request
Related items