Modeling And Optimization Of A Drayage Problem With Foldable Containers

Container drayage transportation is a transportation mode using trucks in a local area before or after marine or railway transportation.Compared with marine transportation and railway transportation,the cost of unit distance and unit container generated by drayage transportation is relatively higher.Container drayage transportation is one of main sources of atmospheric pollutants.Foldable containers can be folded when they are empty.For example,the four-in-one foldable container,whose volume under folding state is only a quarter of it under unfolding state,can save 75%of storage space.The application of foldable containers in drayage transportation is an effective way to save transportation cost.Considering a class of foldable container drayage transportation problem,this thesis develops some research work as follows:(1)Related articles are retrieved.The current research situations about the container drayage transportation problem,the foldable container transportation or empty containers reposition problems and some related algorithms are summarized and analyzed.(2)An assumption that a truck can only transport one type of empty containers or full containers in a planning period(usually one day)is presented.A mathematical description of a class of foldable container drayage problem is given based on the determined activities on vertex(DAOV)graph.There are two types of tasks:collection tasks and distribution tasks.In addition,corresponding activities of empty container transportation are also concerned.The problem is described as two sub problems.The full container sub problem is transferred as a multiple traveling salesman problem with time windows;the empty container sub problem is described as a vehicle routing problem with time windows,where the demand of goods in customers can be positive or negative and the goods can be shared among different customers.Two sub problems have some coupling constraints,such as the total trucks number constraint.Finally,a mixed integer linear programming model is established.(3)An algorithm based on reactive tube search(RTS)is designed for the model established in(2).The algorithm uses a coding structure with double chains,decodes according to the idea of the topological sort,and readjusts visiting times of vertexes.To eliminate the symmetry among encodings,all task fragments visited by different trucks are sorted before a current solution of each iteration is stored.Based on a large number of instances generated randomly,the validity of the algorithm is verified.Results of the RTS algorithm are compared with the results from CPLEX.The comparison with the situation using standard containers shows that using foldable containers can save cost considerably in drayage transportation.(4)The assumption in(2)is relaxed and a more general foldable container drayage transportation problem is described according to a range-based state transition method.In the state transition method,the range instead of the exact value is given to define the state of a truck so that it is more convenient to express the state transition and the state transition time.A mathematical model is established and an algorithm based on RTS algorithm is designed.Two types of instances are generated:scattered customer points instances and clustered customer points instances.Based on a large number of instances generated randomly,the validity of this method is verified.Finally,compared with the situation in(2),some preliminary management revelations have been obtained.