Study On The Mathematical Modeling And Algorithm For Fleet Planning

The fleet planning, as an important decision in the development strategies, determines the survival and the long-term development of the shipping companies. With the constant expansion of the world's shipping scale, importance of the research on the fleet planning is becoming more and more obvious. Through reviewing and summarying lots of literature on fleet planning, we found that the solutions to the large-scale, multi-route, multi-ship fleet planning problems in complex situation is still not very satisfying, relatively little work has been done in the liner fleet planning, and the research on fleet planning under uncertainty has taken the initial step at present. So the research on fleet planning needs to be further improved. Based on the above problems, the main studies in this paper are as follows:(1) A mathematical model of fleet planning with multimode investment in complex market environments is established, which can optimize the decisions on ship investment, updating, routing and fleet size simultaneously. In consideration of the nonlinear influence to the decision-making caused by the ship speed changing on this basis, a nonlinear model for fleet planning is established, and a hybrid particle swarm optimization algorithm is developed to solve the nonlinear model.(2) Two type of fleet planning problems of multi-call liners and trunk line and feeder system are researched. Mathematical models are presented respectively to optimize liner fleet planning problems in the two cases that the same type or variety types of ships are deployed on the same route. Taking the complexity of solving these models into account, the original model is further improved and a Lagrangian relaxation heuristic algorithm is designed.(3) In view of the varying actual operating environments, decision-making problems of fleet planning under uncertainty are researched. by introducing the robust optimization approach based on scenario analysis and by using a scenario set with given probability to describe the uncertainty of market demands, a robust optimization model for fleet planning under demand uncertainty is established. The effectiveness of this robust model is demonstrated by a calculation example. (4) A large shipping company is used as an example to make an empirical analysis. The large-scale, multi-route, multi-ship fleet planning problem in complex situations is solved by using of Benders decomposition algorithm to test and verify the application effect of the new proposed models and algorithms to large-scale practical problems.