Research On Probabilistic Airspace Congestion Management Based Upon Uncertain Traffic Demand Prediction

With the rapid development of air transportation industry, airspace system performance presstureis increasing, and the airspace environment has been more and more complicated. Because uncertainfactors have increasingly been affacting the air traffic, how to releivate the airspace system pressturefrom uncertainty theory perspective is becoming a focus in the field of air traffic management amongsome advanced aviation counties. In order sophisticat the process of airspace congestion problem inChina, based on the current situation of airspace system, we deeply analyzed the random changemechamism of air traffic demand. Meanwhile, we investigate the airspace congestion managementfrom the point of risk management, and discussed how the traditional determinated airspacecongestion management gradually transferred to the uncertain management. The research is help toconsummate the method of air traffic demand prediction, to improve the theory of airtspacecongestion management, and promote the economic and social benifitIn this paper, based on some abroad sophisticated experiences, we researched the praticalairspace congestion problem. We utilized the thecnologies in the field of air traffic flow management,air traffic flow stasatics and prediction, multi-objective optimization and airspace congestion riskmanagement, and discussed the process of uncertain air traffic demand prediction, and interegratedthe risk prediction, resolution and decision into the airspace congestion relievation. The main contentof the paper is as follows:(1) Reaserach base of airspace congestion risk management was introduced. We defined the basicconcepts including air traffic management, air traffic flow management, risk management, risk trafficprediction, uncertain airspace capacity, airspace congestion risk and airspace congestion riskmanagement, and also established the deviation relationship between these concepts. Then, weanalyzed the basic theory of airspace congestion management, and provided the probabilistic airspacecongestion management definition, and its specific process and inherent. Moreover, based on theconcepts of probabilisitic airspace congestion management, we developed the relationship betweenthe concepts and the technologies, and introduced the technologies including air traffic flowmanagement, risk management, air traffic flow statictics and prediction, airspace capacity evalution,multi-objective optimization, and airspace congestion risk management.(2) Methodology of airspace congestion risk prediction was studied. First, we exacted the mainfactors influenceing the airspace sector’s traffic demand prediction, and established the airspace sectorprobabilistic demand prediction model through analyzing the random features of the aircraft sector entry time, sector exit time and existing time in the airspace sectors. Based on the model, the airtraffic demand probabilistic distribution was obtained, and the prediction errors could be quantified.Then, based on the airspace sector probabilistic traffic demand prediction model, we combined certainand uncertain airspace capacity ecaluation respectively, and established the airspace congestion riskprediction model and method. The simulation results show that the model and method we establishedcan quantify the traffic demand prediction uncertainty and its changing rules, and the influences of theuncertain factors on the traffic demands and capacities can be found out.(3) Methodology of airspace congestion risk resolution based on partial optimization wasinvestigated. Based on the airspace congestion risk prediction, we established the airspace congestionresolution model. Through this model, we could comprehensively deal with several objectives,including the follow-up congestion, the total flight delay time, the different airspace users’ allocationequity and influence of the resolution strategies. Meanwhile, we designed high-dimensionalmultiobjective genetic NSGA2, realizing the initial optimization of the airspace congestion resolutionstrategies. The simulation results show that the model and algorithm established cannot only find outoptimum departure time and routes, but also reduce the risk of follow-up congestion, the operationcost and the influence, meanwhile, increase the equity of the airspace users.(4) Methodology of airspace congestion risk resolution based on overal optimization wasresearched. First, in order to deal with the remaining problem of airspace congestion globaloptimization, we established the airspace congestion risk decision model, considering the distributionand balance of the air traffic flow, the workload of the air traffic controlors. Meanwhile, usingmulti-objective genetic algorithm, we reoptimized the airspace congestion risk resolution strategies.Combined with the airspace congestion resolution, we developed the airspace congestion riskevulation model, and utilized the decision tree to resolve the operating time of the airspace congestionresolution strategies and the threshold of the airspace congestion probability. Finally, the simulationresults show that the model and method established can deal with the air traffic global distribution andthe operation workload in each sector, and can obtain the suitable operating time and the airspacecongestion risk threshold.Finally, in the thesis, we concluded the archievements of the probabilistic airspace congestionmanagement based on the uncertain traffic demand prediction, and pointed out the deficiency of thisresearch about microscopic traffic flow, human factors and performance system relization. Accordingto the conlusion above, we suggested the future study direction.