【作者】 唐小卫；

【导师】 朱金福；

【作者基本信息】 南京航空航天大学 ， 交通运输规划与管理， 2009， 博士

【摘要】 由不正常航班引发的问题不仅给旅客带来诸多不便,也给航空运输系统造成巨大的经济损失。不正常航班问题已经成为影响航空运输系统运营效率和服务质量的核心问题。航空运输系统由航空公司、机场和空管三个子系统组成,协同决策(Collaborative Decision Making,CDM)机制已被欧美等航空发达国家证明是提高航空运输系统效率的有效管理技术。它是一种协同合作的理念,通过整合空域管理者(空管)和使用者(机场和航空公司)提供的数据得出更准确的信息,并且使管理者和和使用者共享决策信息。研究协同决策机制下航空运输系统不正常航班问题,对于提高航空运输系统的竞争力,满足日益增长的航空运输需求,具有重要的意义。本文以协同决策机制下航空运输系统不正常航班相关问题为研究对象,主要研究工作如下:构建了CDM机制下不正常航班问题决策系统框架,并详细阐述框架内各子系统的决策主体、信息需求、信息传递方式和平台支持,分析了框架内各子系统的决策模式、决策要求和决策内容等。CDM机制要求各航空公司及时制定出最优的不正常航班恢复策略,这是各航空公司和空管信息交流的基础。为此首先研究了航空公司不正常航班恢复的建模问题,详细分析了已有的资源指派模型、多商品网络流模型和时间离散近似模型的优点和不足,在此基础上将飞机就绪时间、飞机流不平衡约束、航班时空衔接特性、多机型调度和合并航班成本计算等问题加入资源指派模型,使新模型更符合CDM机制下航空公司的实际运行。然后针对新建的不正常航班恢复模型设计了一种启发式算法——贪婪随机模拟退火算法。新算法通过限制飞机路线对的数量以提高算法的时间效率;在GRASP算法的基础上添加航班串取消操作作为邻域解的构造方式之一,以解决流不平衡;设计邻域解备选池并引入基于模拟退火算法的搜索策略,有效降低了问题陷入局部最优解的概率。实例表明所建模型和设计算法的有效性,其结果与GRASP算法相比具有明显的优越性,在算法的时间效率和求解质量之间取得较好的平衡,能够满足CDM机制对航空公司实时决策和优化决策的要求。机场终端已成为空中交通网络的瓶颈,现有的研究工作重点关注飞机正在执行的航班性质,忽略了前后航班的衔接关系和延误波及问题。本文基于CDM机制提出将航空公司航班优先选择权引入机场终端流量优化模型,并分析航班取消信息和航班优先选择信息对流量优化方案的影响。机场场面管理是CDM的应用领域之一,飞机推出是机场场面管理的重要内容。为解决不正常航班影响下飞机推出地面冲突增多、航班延误加剧和机场盲区影响等问题,分析了飞机推出流程,设计了飞机推出程序,包括推出程序命名、停机位组合和推出程序标识等内容;建立了停机位组合优化设计模型,提出一种层级迭代搜索算法,并实例验证了模型和算法的有效性。对飞机推出程序进行三维仿真研究。在考虑了飞机滑入对飞机推出的影响后,选取模糊随机Petri网作为建模工具,使推出模型具有模糊随机性;以ServiceModel软件为平台,开发飞机推出三维可视化仿真系统,重点对推出延迟时间、推出安全间隔和单位时间推出架次三项指标进行分析。仿真结果表明设计的推出程序方案在安全和效率方面均能满足停机坪运行的要求。更多还原

【Abstract】 Problems arising from irregular flights not only bring inconvenience to passengers, but also cause tremendous economic loss to air transportation system. Irregular flights have become the core problem affecting the operating efficiency and service quality of air transportation system. Air transportation system consists of three subsystems, namely airlines, airport and air traffic control. Collaborative Decision Making (CDM) has been proven to be the effective management technique to improve the efficiency of air transportation system by aviation developed countries in Europe and America. CDM reflects a collaborative concept, which manages to obtain more accurate information through integrating the data provided by airspace manager (air traffic control) and the users (airport and airlines) so that the manager and users could share decision-making information. The research on irregular flights recovery in the air transportation system under CDM is of important significance to improve the competitiveness of air transportation system and to meet the increasing demand of air transportation.This article uses relevant problems about irregular flights in the air transportation system under CDM as research objects, and the main research work is as follows:This article builds the decision-making scheme frame of irregular flights problems in the air transportation system under CDM, and describes in detail the decision maker, information need, systematic platform support, information transmission, decision-making mode, decision-making requirement and decision-making contents of the frame.CDM requires airlines to make the optimal irregular flights recovery decision in time, which is the basis of information exchange between airlines and air traffic control. Therefore, this article first studies the modeling of irregular flights recovery of airlines, and analyzes in detail the strongpoint and deficiency of the existing Resource Assignment Model, Multi-commodity Network Flow Model and Time Band Approximation Network, and on this basis, questions such as aircraft ready time, aircraft imbalance constraint, flight connectivity, multi-type aircraft allocation and cost calculation of combined flight are added into the amended Resource Assignment Model, so that the new model is more suitable for the practical operation of airlines under CDM.Then, a heuristic algorithm named greedy random simulated annealing algorithm is designed for the new irregular flights recovery model. The new algorithm improves the time efficiency of the algorithm by restricting the number of aircraft route pairs. On the basis of GRASP (greedy random adaptive search procedure), cancelling flight string is added as one of the composition method of neighborhood to solve aircraft imbalance. Neighbors list is designed and search strategy based on simulated annealing algorithm is introduced so as to effectively reduce the probability that questions may fall into local optimal solution. Examples prove the effectiveness of the new model and designed algorithm, whose result has obvious advantage over the GRASP algorithm. The designed algorithm has achieved a comparatively good balance between time efficiency and solution quality, so that the requirements on real-time decision-making and optimal decision-making of airlines put forward by CDM could be met.Airport terminal has become the bottleneck of air traffic network. The existing research work pays attention to arrival and departure flights, but neglects the cohesive relationship between the two successive flights and the problem of delay propagation. On the basis of CDM, this article proposes to add First Priority of Flight to the airport terminal flow optimization model, and analyzes the influences of flight cancellation information and flight priority information on flow optimization scheme.Airport surface management is one of the application fields of CDM, and aircraft pushback is an important content of airport surface management. In order to solve the problems such as increasing conflicts of aircraft pushback, increasingly serious flight delay and influence of airport blind area under the influence of irregular flights, this article analyzes aircraft pushback flow and designs the aircraft pushback procedure of busy airports, including pushback procedure naming, gate position combination and pushback procedure signage. This article also builds optimal design model of gate position combination, puts forward a hierarchy iteration algorithm, and proves the effectiveness of the model and algorithm by examples.The 3D-Simulation research on aircraft pushback procedure is conducted. Considering the relationship between the aircraft taxi-in and pushback, fuzzy Stochastic Petri Nets are chosen as modeling tool so as to ensure the randomicity of pushback model. Aircraft pushback 3D visible simulation system is developed by using ServiceModel software as platform. Three items of indices, namely pushback delay time, pushback safe distance and pushback aircraft number per unit time are analyzed. The simulation result indicates that the designed pushback procedure could meet apron operation needs in both aspects of safety and efficiency.更多还原