Adaptive Genetic Based Crew Rostering And Its Application

The crew scheduling is a key sector in process of airline producing. Therefore, the scientific and reasonable scheduling of crew would be useful to make the best of human resources, to reduce operating cost and to improve marketable competitive power. The crew scheduling belongs to the typical combination-optimization problem and has some traits such as complexity, huge data, multi-aims and multi-restriction. It is usually solved by subsection.The paper firstly describes the flow and traits of the crew scheduling, then analyzes and compares three modern heuristic methods, and according to its specialties, focuses on the adaptive genetic algorithm and researches on it intensively. Aiming at the content of the crew rostering, it builds a mathematical model which makes the equilibrium of task and line's difference be key aims, and makes others be restrictions. With that the paper discusses that how the AGA solves the problem concretely. Following, it points out that the crew scheduling is realized in a system called Flight Optimization & Revenue Maximization and analyzes the adaptation, transparence and distensibility of the system. In addition, this paper discusses some problems about producing the original feasible solution, and improving the probability of selection and operation of crossover, in order to provide a way towards greater efficiency.