Research On Some Vehicle Routing Problems With Backhauls

The vehicle routing problem with backhauls (VRPB) is an extension to the vehicle routing problem (VRP) which serves either linehaul customers or backhaul customers. VRPB can serve different type of customers by a fleet of vehicles.Classical VRPB have the restriction that all backhauls have to be visited after all the linehauls, it will make the route bypass and can not solve the customers where have both delivery and pick-up demands. These years, vehicles have two more entrances since the improvement of the technology. Lots of papers about VRPB neglect the rearrangement of the goods, and regard the real capacities of vehicles as the restriction of the problem. But many real problems can not be solved except for the cargo have the same figures. In this paper, we improved VRPB and allow mixed linehaul and backhaul customers along the routes which can avoid the rearrangement of the goods.In this paper, we will summarize the study on the VRPB, and then we will introduce the VRPB in particular, which is taking into four parts.1. The simple improved VRPB Firstly we give a new parameter to improve the VRPB, then we can ensure the route least and avoid the rearrangement of the goods; secondly, we use the improved GAs to solve the improved model, and design a new process of feasible mutation to ensure the feasibility and avoid the part-best solution.2. VRPB with time windows. We add a restriction of time window on the simple improved VRPB, and then the model is more fitful to the actuality of logistic transportation. We use the branch and bound, GAs and integral programming to solve the problem. At last, the instances for problem prove its validity.3. VRPB with fuzzy variable. The study about the above two parts is on the base of exact parameters, but in the real world, the parameters are also fuzzy, for example, you also see "almost", "between" in the parameters. So in this part, we will discuss the method of solving the fuzzy problem.4. VRPB which is dynamic and random. The study of the above three parts is on the known of all the parameters of the customers. In this part, we will discuss the dynamic and random problem, and use the method of labels and inserting to solve the...