详情

当前位置：首页> 国外交通科技报告数据库 >详情

原文传递 Large-Scale Network Design Using Composite Variables: An Application to Air Mobility Command's 30-Day Channel Route Network

题名：	Large-Scale Network Design Using Composite Variables: An Application to Air Mobility Command's 30-Day Channel Route Network
作者：	Nielsen, Christopher A.;
关键词：	LINEAR PROGRAMMING, ROUTING, AIR TRANSPORTATION, LOGISTICS PLANNING, MOBILITY, DECISION MAKING, NETWORKS, THESES, VARIABLES, SCHEDULING, MISSIONS, CARGO, AIRMOBILE OPERATIONS.
摘要：	Each month, the United States Air Force's Air Mobility Command is responsible for designing a large-scale air mobility network - called the channel route network that is used to transport military personnel and cargo throughout the world. Traditionally, planning the channel route network has been a manual process, requiring numerous hours to generate a monthly channel route schedule. We formulate the monthly channel route scheduling problem using traditional network design formulations, and we use Dantzig-Wolfe decomposition to overcome tractability issues. The resulting linear programming bounds on the optimal integer solution are weak, however. Furthermore, the formulation is unable to capture complex operational rules and regulations that govern aircraft flows and cargo flows through the channel route network. To overcome these challenges, we apply a variable redefinition technique known as composite variable modeling. Using this technique, we alleviate the need to explicitly include cargo flow decisions by implicitly capturing them in the design variables, which are then combined to create composite variables. A single composite variable represents the selection of one or more aircraft missions that completely cover a subset of cargo commodities. The resulting formulation is computationally superior to traditional network design formulations because it achieves tighter bounds - allowing excellent integer solutions to be found quickly - and it is able to capture complex operational rules and regulations. We illustrate that the nature of the formulation can be easily changed depending on the desired output by simply modifying the composite variable generation procedures, which does not compromise the structural properties of the formulation that allow it to achieve strong linear programming relaxations and short solution times.
总页数：	197
报告类型：	科技报告