Research On Intelligent Evaluation System Of Shenhua Railway Scheduling Day Shift Plan

In order to improve the efficiency of Shenhua railway transportation organization and dispatching command,promote the automatic and intelligent preparation of dispatching day shift plans,realize timely analysis and feedback of dispatching day shift plans,and realize the transformation of Shenhua railway transportation organization mode from dispatching and commanding to plan-oriented for a fundamental change,Shenhua Railway urgently needs to build a complete and efficient system of intelligent evaluation for dispatching day shift plans as the evaluation module support for the intelligent preparation system of dispatching day shift plans.The intelligent evaluation system of Shenhua railway dispatching day shift plan includes two parts: evaluation index system and intelligent evaluation model.With the decision-making model,through the "indicator-evaluation-decision" layer-by-layer progression,this paper realizes the timely-automated-intelligent evaluation and decision-making of the new preparation plan,and achieves timely feedback to the scheduling decision-maker and the intelligent preparation system.Based on system composition and extension of intelligent evaluation,the main research work of this paper is as follows:(1)The evaluation index system of Shenhua railway dispatching day shift plan is constructed.Based on the analysis of Shenhua Railway dispatching day shift planning business and the necessity of intelligent evaluation,the evaluation objectives are clarified,and the index system construction process and index selection principles are explained.From the two perspectives of the quality evaluation index of dispatching day shift plan and the environmental variable index,the evaluation index system of Shenhua railway dispatching day shift plan is constructed through index clustering and screening.(2)The intelligent evaluation model of Shenhua railway dispatching day shift plan is constructed.This paper clarifies the basic concepts of intelligent evaluation,compares and analyzes the applicability of different evaluation models,taking the Bayesopt theory and the Gaussian process regression(GPR)machine learning model as the basis,improves the the training performance of GPR model through the Bayesopt parameter adjustment method.On the basis of these,an intelligent evaluation model modeling method based on Bayesopt-GPR is proposed,and the intelligent evaluation model is constructed according to the actual data.Through the learning of historical evaluation results,the automatic evaluation of new samples is achieved on the premise of ensuring that the evaluation results conform to empirical cognition.(3)Three-decision-making models of Shenhua Railway dispatching day shift plan is constructed.The basic model principle of the three-decision-making rough sets is clarified based on the basic theory of rough sets,also,the calculation method of the probability threshold and conditional probability are explained.Further more,a decision-making model for scheduling day shift planning based on the three-decision-making rough sets is constructed,and the "evaluation-decision" linkage mechanism is given.Based on the intelligent evaluation results,the decision-making model supports the application of the intelligent evaluation system.(4)Example verification of the intelligent evaluation system of Shenhua railway dispatching day shift plan is given.The overall framework and workflow of the intelligent evaluation system are given.By taking Shenhua Railway Baoshen Railway Group as an example,the feasibility of the intelligent evaluation system is verified,and the applicability of different machine learning models is compared,also,the advantages of intelligent evaluation compared with traditional evaluation are illustrated,finally,the decision analysis of the day shift plan is carried out based on the intelligent evaluation results.