Models And Algorithms Of The Uncertain Transportation Problem

Along with appearance of the importance of logistics and supply chain in the global economics, the importance of material dispatching is emerging increasingly, which is the essential part of logistics. Transportation models play an important role in logistics and supply chain management for reducing cost and improving service.Traditional transportation problem, that is to say, the transportation problem in the certain environment has been studied very well. However, in the actual life, for the uncertainty of weather, road structure or traffic, as a result, the transportation problem in the uncertain environments became very important. Although there are many scholars investigate it, it is very few to apply uncertain programming to transportation problem. So, it is very necessary to construct uncertain parameters transportation models and solve it.This thesis based on the uncertainty theory: probability, fuzziness, rough set, from the appearance of uncertainty--- randomness, fuzziness, roughness, together with the uncertain programming technique, and then systematically and roundly researched on the math ideology, math model, character of model and arithmetic of the uncertain multi-objective transportation problem.Now we shall list our contribution to the transportation problem:(1)Expected value goal programming, chance-constrained goal programming model and dependent-chanced goal programming model of random transportation are constructed ;(2) Expected value goal programming, chance-constrained goal programming model and dependent-chanced goal programming model of fuzzy transportation are constructed;(3)Expected value goal programming, chance constrained goal programming model and dependent-chance goal programming model of rough transportation problem are constructed;(4)enlightened by Liu's thought that solve the uncertain programming with hybrid intelligent algorithm, we design a hybrid intelligent algorithm, that is, genetic algorithms based on simulation to achieve the approximate best solution of the nine above mentioned models. Then, examples are listed to verify the feasibility and efficiency of our algorithm.