铁路车站取送车作业问题一般模型解的构造方法

A Method for Generating Solutions of the Generic Model for Taking-Out and Placing-In Wagons in Railway Station

基于取送车作业问题具有顺序和批次2个优化维度,根据客观约束条件和实际生产经验,设计取送车作业问题一般模型解的模块化构造方法。以满足取送车作业顺序偏序约束构造取送车作业初始顺序方案,以满足装卸区容车能力约束构造取送车作业调整顺序方案。为了满足相应放射枝不同批次约束、调机牵引能力约束、时间窗约束和提高调机运用效率,分别采用基于相异放射枝不同批次约束的批次划分方法、基于牵引能力最大化的批次划分方法、基于不必要牵引车组数最小化的批次划分方法、基于及时送车原则的批次划分方法和基于及时取车原则的批次划分方法,将取送车作业调整顺序方案划分为一系列有序的取送车批次。算例仿真结果表明:利用本文方法生成的取送车作业方案满足铁路车站取送车作业问题一般模型的约束条件,使调机的运用和各装卸区的集中作业强度更为合理,使有效方案的生成效率比随机生成方法更高。

According to the characteristics that wagon's taking-out and placing-in problem has two optimization dimensions of sequence and batch, a method used to generate solutions of generic model for wagons' taking-out and placing-in problem in railway station is proposed based on some objective constraints and practical experiences. The initial and adjusted sequence schemes for taking-out and placing-in operations are successively generated to satisfy the partial order relation of taking-out and placing-in operations and the wagon-accommodation capacity in loading and unloading zones. To meet some constraints including operations for different radiation branches included in different batches, traction capacity of shunting locomotive and time windows and improve the operation efficiency of shunting locomotive, the adjusted sequence scheme is divided into a series of ordered batches by a set of methods which are respectively based on operations for different radiation branches included in different batches, the traction capacity ability maximization of shunting locomotive, the unnecessary taking-out and placing-in operation minimization of shunting locomotive, the timely wagons' placing-in and the timely wagons' taking-out. An example simulation shows that the obtained operation schemes meet the constraints for wagons' taking-out and placing-in problem in railway station and make the operation intensity in loading and unloading zones and the shunting locomotive application more reasonable, and the proposed methods make the operation scheme generation more efficient than random ones.