یک ابزار پشتیبانی تصمیم گیری یکپارچه برای برنامه بهبود در طول عملیات های نامنظم شرکت های هواپیمایی

This paper presents a decision support tool for airlines schedule recovery during irregular operations. The tool provides airlines control centers with the capability to develop a proactive schedule recovery plan that integrates all flight resources. A rolling horizon modeling framework, which integrates a schedule simulation model and a resource assignment optimization model, is adopted for this tool. The schedule simulation model projects the list of disrupted flights in the system as function of the severity of anticipated disruptions. The optimization model examines possible resource swapping and flight re-quoting to generate an efficient schedule recovery plan that minimizes flight delays and cancellations. A detailed example that illustrates the application of the tool to recover the schedule of a major US air-carrier during a hypothetical ground delay program scenario is presented. The results of several experiments that illustrates overall model performance in terms of solution quality and computation experience are also given.

In this paper, a decision support tool for airlines schedule recovery during irregular operations is presented. The tool allows airlines controllers to detect schedule disruptions ahead of their occurrence and to generate an integrated recovery plan for all used resources. A rolling horizon modeling framework with greedy optimization strategy is adopted. The framework integrates a schedule simulation model and a resource assignment optimization model. The flight simulation model predicts the list of disrupted flights in the system as a function of the severity of anticipated disruptions. The optimization model combines different recovery actions in one efficient plan to minimize projected flight delays and cancellations. A detailed example that illustrates the application of the tool to recover the schedule of a major US air-carrier during a hypothetical ground delay program scenario is presented. The results of several experiments that illustrates overall model performance in terms of solution quality and computation experience are also given. Based on these results, the tool was capable to generate an efficient recovery plan with considerable flight savings. Furthermore, all recovery plans are generated in less than 1 minute, which allows near real-time schedule recovery. Several extensions are considered for this research work. First, through considering flight landing slots as one of the resources to be assigned for each flight, DSTAR could be extended to provide schedulers with an efficient slot allocation scheme as part of the Collaborative Decision Making (CDM) process adopted by the FAA. The resulting slot allocation scheme will ensure minimum system-wide disruptions. Furthermore, if information of airport gate availability is accessible, this information could be incorporated in the generated recovery plan. Developing a hybrid optimization strategy in which greedy resources assignment is modified to include look-ahead optimization capabilities is another extension. For instance, if an aircraft is scheduled for maintenance at some time in the future and this maintenance activity could be conducted in several stations. Feasible aircraft routes that end at all these stations could be generated. An aircraft assignment within a stage is marked feasible if it coincides along one of the generated routes. Thus, it ensures that local decisions made at each stage are resulting in a solution that satisfies maintenance constraints beyond the limits of this recovery stage.