| Fleet Assignment is one of the most important parts of the management of Airline Company. The result of the assignment directly affects the profit of the company. Therefore, how to automate the fleet assignment using appropriate algorithm and solving the fleet assignment problem more reasonable and effectively comes to be an important job of domestic airline to enhance competition and to control cost.This thesis introduces Discrete Particle Swarm Optimizer (DPSO) Algorithm to solve the fleet assignment problem. Firstly, the characterization of the fleet assignment is analyzed. Then it presents the basic mathematical model of the fleet assignment problem and some other mathematical models with different constraints, and especially the realization technology of fleet assignment. Then the thesis analyzes several typical algorithms which have been used to solve the fleet assignment problem and contrast them with each other. In consideration of the data size and the constraints of fleet assignment problem, we choose the DPSO algorithm which has been used to solve many Combinatorial Optimization problems and was proved to be effectively. Firstly, we define the parameters and the operating principles of DPSO algorithm in view of the attributes and constraints of the fleet assignment problem. Then we make use of the flight-swapping operation to generate the speed of particle in every dimension, and introduce the repulsion operator to make the particle more various and keep the evolutionary ability of the particle. After these definitions, we build the mathematical model of the fleet assignment system using DPSO algorithm and realize the system. Then we present the main frame of the system and introduce the realization and design of every module.In the end of the dissertation, we analyze and assess the performance of the system and give some further suggestions to improving the system. |
