Parametric model for evaluating railway network capacity using neural network techniques

Abstract

Railway capacity is a significant indicator of the railway networks performance. The capacity will be evaluated using one of the modeling techniques that consist of large numbers of variable factors with relationships between each other that affect the railway capacity. This research uses neural network as a modeling technique, which is developed using an artificial intelligence that offers an enhanced approach to estimate capacity. The proposed model can provide government and railway agencies with valuable information about factors that influence railway capacity and can be used to evaluate t he capacity of railway under different scenarios to improve its performance. The model is based on the official timetable representing the actual operation conditions. A case study is applied on Egyptian Railway Network and validated by determining the root mean square error and the relative error. The maximum capacity is estimated under different conditions such as track, signal and traffic conditions. Thus, the proposed study rearranges the given factors according to their importance.