Modeling, Analysis And Simulation Of Air Traffic Network Based On Hybrid Multi-agent System Theory

The rapid development of the Chinese civil aviation industry results in a sharpincrease of the flight number. Meanwhile, the limited air space resource and thelagged techniques of the transportation management lead to severe un-safety andflight delay. Aiming at solving these problems, researchers focus on the studies of thecomplexity of the airspace, the process of the air traffic control, the management ofthe airlines, the airport regulation etc, and try to improve the air transportationmanagement.The air traffic control process within the limited airspace structure is the mostimportance factor that affects the flight delay problem. Therefore, based on themodeling and the analyzing of the whole air traffic network, this thesis proposes aproject to measure the real-time workload of the controller, which provides thefoundation for the division of the airspace sector and the dynamic allocation of theairlines. The research is conducted in the following aspects:First of all, from the whole air traffic system level, according to the hybridfeature of the continuous flight process of the aircrafts and the discrete control processof the controllers, the hybrid system theory is introduced into the air traffic system toconstruct the model of the air traffic network, including the flight process modeling ofthe single aircraft based on the hybrid automaton model and the conflict resolutionmodeling of the multi-aircrafts.Secondly, in the conflict resolution process, the mid-term resolution affects thewhole traffic efficiency most, and the widely used genetic algorithm ha s lowtimeliness and may easily fall into the local minimum, these drawbacks limit theapplication of the genetic algorithm in the conflict resolution. Thus a heuristicalgorithm with constraint is proposed, and comparison analyses are conducted for theresolution process with crossover conflict. The result shows that the proposedalgorithm could achieve the correct resolution process, and elapsed time is shorter. Inorder to show the conflict resolution process vividly, a simulation platform isconstructed by using MATALB to conduct the virtual reality simulation for theconflict resolution process of the aircraft.Thirdly, on the basis of the air traffic model, the thesis analyzes the quantitativerelationship between the complexity of the airspace and the control actions. Thenaccording to it a method is proposed to calculate the real-time workload of thecontroller based on the airspace complexity factors. The results show that, whencompared with the method based on the aircraft number, the proposed method couldachieve a more accurate workload result of the controllers. Moreover, the method is easy to carry out and thus is suitable for the real-time workload measurement of thecontrollers in the traffic process, which provides the foundation of the airspace sectordivision and the dynamic allocation of the airlines.Finally, an air traffic simulation system is constructed on the Vc++6.0platformbased on the Map component MapX. This system could upload the real staticstructure information of the airspace and conduct the deployment of the flightscheduling and the control action，which could realize the dynamic simulation of thereal air traffic system, the online workload measurement of the controllers and thedynamic division of the airspace sectors. The simulation result for the airspace ofnorthern China could provide some reference for the real performance analyses anddecisions of the air traffic system.The thesis research constructed a model for the air traffic system based on thesystem level and utilized the hybrid system theory to analyze the conflict resolutionand the workload calculation problem, which provided a new solution for themodeling and analyzing of the air traffic system. The follow-up research is suggestedto focus on each module of the system, including the dynamic airspace sector division,the dynamic allocation of airlines and etc.