详情

原文传递 Optimal Planning of Rail Grinding Activities in Large-scale Networks.

题名：	Optimal Planning of Rail Grinding Activities in Large-scale Networks.
作者：	Lei, C.; Ouyang, Y.; Xie, S.
关键词：	Rail grinding activities, Rail transportation
摘要：	In railroads, safety is one of the most important topics and has attracted tremendous attention recently due to some reported accidents. Among all the causes of train accidents in the United States, track defects are one of the leading reasons. Basically, the natural processes of wear and fatigue of rail steel can proceed at a rapid pace that results in track defects and short service lives. Therefore, maintenance of rails (such as grinding, ballast cleaning, ditching) is very important for regular railroad operations. One unique feature of these rail maintenance activities is that vehicles may experience variable productivity due to both endogenous and exogenous factors, and thus the actual working duration of a maintenance job is unknown to the operator beforehand and needs to be determined. For example, if a track has not been grinded in a timely fashion, its abrasive condition will further accelerate the wear and tear such that more time has to be spent on that track next time to get the job done. The variability of productivity in this case is caused endogenously by the generated grinding schedule. On the other hand, vehicle productivity may also be affected by exogenous reasons such as traffic interference. Take ballast cleaning as an example, the associated vehicles can work for a longer time during a day if the track to be maintained is under traffic curfew. Practical instances of the rail maintenance routing and scheduling problem (RMRSP) usually involve hundreds or thousands of jobs and an enormous number of complex constraints. Meanwhile, all the routing decisions must be made in a large-scale railroad network. In current rail industry practice, routing and scheduling decisions are mostly manually determined based on the experiences and knowledge of experts. However, such decision-making process is likely to take a long time while the solution quality may still be poor. In light of these complexities, this research focuses on building a comprehensive mathematical model and developing efficient solution approaches for RMRSP under variable productivities.
报告类型：	科技报告