计算机工程与应用
計算機工程與應用
계산궤공정여응용
COMPUTER ENGINEERING AND APPLICATIONS
2013年
14期
19-24
,共6页
李奔驰%党创寅%郑津津
李奔馳%黨創寅%鄭津津
리분치%당창인%정진진
航空扰动管理%不正常调度%整数规划%分布式计算%字典序%消息传递接口(MPI)%OpenMP
航空擾動管理%不正常調度%整數規劃%分佈式計算%字典序%消息傳遞接口(MPI)%OpenMP
항공우동관리%불정상조도%정수규화%분포식계산%자전서%소식전체접구(MPI)%OpenMP
airline disruption management%irregular operation%integer programming%distributed computation%lexicographical order%Message Passing Interface(MPI)%OpenMP
在航空公司的运作中时常会出现干扰它正常运作的现象。在这种情况下,航空公司必须马上制定航线修复计划使受到干扰的航线尽快复原,以防止更大面积的航班取消和航班延误。提出一种基于递增映射迭代方法的分布式整数规划算法来解决由于机场关闭引起的航线扰动问题。整个问题分成了两个子问题:可行航线的生成和飞机的重指派。第一个子问题的问题空间被初始点分割方法分割成了若干片段。然后在一个分布式的计算网络中使用递增映射迭代方法在分得的每个片段上同时求解第一个子问题。得到的可行航线用来求解第二个子问题。最后的算例结果可以发现提出的方法要好于CPLEX和多目标基因算法。
在航空公司的運作中時常會齣現榦擾它正常運作的現象。在這種情況下,航空公司必鬚馬上製定航線脩複計劃使受到榦擾的航線儘快複原,以防止更大麵積的航班取消和航班延誤。提齣一種基于遞增映射迭代方法的分佈式整數規劃算法來解決由于機場關閉引起的航線擾動問題。整箇問題分成瞭兩箇子問題:可行航線的生成和飛機的重指派。第一箇子問題的問題空間被初始點分割方法分割成瞭若榦片段。然後在一箇分佈式的計算網絡中使用遞增映射迭代方法在分得的每箇片段上同時求解第一箇子問題。得到的可行航線用來求解第二箇子問題。最後的算例結果可以髮現提齣的方法要好于CPLEX和多目標基因算法。
재항공공사적운작중시상회출현간우타정상운작적현상。재저충정황하,항공공사필수마상제정항선수복계화사수도간우적항선진쾌복원,이방지경대면적적항반취소화항반연오。제출일충기우체증영사질대방법적분포식정수규화산법래해결유우궤장관폐인기적항선우동문제。정개문제분성료량개자문제:가행항선적생성화비궤적중지파。제일개자문제적문제공간피초시점분할방법분할성료약간편단。연후재일개분포식적계산망락중사용체증영사질대방법재분득적매개편단상동시구해제일개자문제。득도적가행항선용래구해제이개자문제。최후적산례결과가이발현제출적방법요호우CPLEX화다목표기인산법。
The operation of an airline schedule often deviates from the original planned schedule due to some disruptions hap-pen. In this situation, the airline must be able to get back to the original schedule by producing recovery plan as quickly as possi-ble. This paper proposes a distributed implementation of an iterative method for integer programming to solve the airline disrup-tion problems caused by the airport closures. The problem is formulated as two sub problems:feasible flight routes generation and aircraft reassignment. The solution space of the first sub problem is divided into several segments by the initial seeds cluster division method. The iterative method is applied to solve the first sub problem in each segment simultaneously in a distributed computation network, and the resulting feasible flight routes are used to construct an aircraft reassignment which is the second sub problem. Numerical results show that the proposed approach performs better than CPLEX and multi-objective genetic algorithm.