Robust runway scheduling under uncertain conditions

详细信息    查看全文

• 作者：Andreas Heidt^a ; ^{andreas.heidt@fau.de" class="auth_mail" title="E-mail the corresponding author} ; Hartmut Helmke^b ; ^{hartmut.helmke@dlr.de" class="auth_mail" title="E-mail the corresponding author} ; Manu Kapolke^a ; ^{manu.kapolke@fau.de" class="auth_mail" title="E-mail the corresponding author} ; Frauke Liers^a ; ^{frauke.liers@fau.de" class="auth_mail" title="E-mail the corresponding author} ; Alexander Martin^a ; ^{alexander.martin@fau.de" class="auth_mail" title="E-mail the corresponding author}
• 关键词：Scheduling ; Uncertainty ; Time-indexed model ; MIP ; Mixed-integer programming ; Dynamic time-indexed model ; Strict robustness ; Light robustness
• 刊名：Journal of Air Transport Management
• 出版年：2016
• 出版时间：September 2016
• 年：2016
• 卷：56
• 期：part_PA
• 页码：28-37
• 全文大小：671 K

文摘

The runway is the main element that combines airside and groundside of the ATM System. Thus, it is crucial to develop efficient models and planning algorithms for its effective usage. The best planning algorithm, however, is useless if the resulting plans cannot be implemented in the real world. This often happens because the input data of the planning algorithms face disturbances or changes over time, respectively. For example, an estimated time of arrival/departure of an aircraft may be changed. It is usually not certain for the next ten hours.

In this work, we study the runway scheduling problem under uncertain conditions. First, we present mathematical optimization models that ignore uncertainties. In the most effective approach, we compute for every discretized point in time whether an aircraft is scheduled and if so, which one is. Then, in each planning step we take uncertainties into account. We then apply different robust optimization methods in order to devise solution approaches that lead to stable plans. These optimization approaches are integrated into a simulation tool and evaluated in different traffic scenarios.

The Monte-Carlo simulations for a mixed-mode runway system show that our robust approaches result in fewer sequence changes and target time updates, when compared to the usual approach in which the plan is simply updated in case of infeasibility. Thus, we show that protection against uncertainties by using robust optimization indeed leads to considerably more stable plans.