According to the serious contradiction between maintenance workload and maintenance time for the ballast track geometry irregularity of the exiting railway,the annual comprehensive maintenance arrangement method of track geometry irregularity using large-scale maintenance machines is analyzed and the corresponding decision-making system is established in terms of the theory of ballast track geometry irregularity together with the performance and character of large-scale maintenance machines and inspection data of track geometry irregularity by track inspection car.The main research work and conclusion in this dissertation involves the follows.(1) The track is divided into many lots,which are short enough(100 m for this research),as the basic management unit for evaluating geometry irregularity,so the standard deviation of track irregularity in each lot as the index of geometry irregularity.Unit is composed of several lots as the maintenance place for large-scale maintenance machines whose maximum length is the maximum possible maintenance length of machines in "skylight" of a day.A period of ten days is the time unit of maintenance machines.The deterioration quantity and the deterioration speed,which describe the deterioration degree of standard deviation of track geometry irregularity in each lot,are calculated.The restoration quantity is calculated by analyzing restoration degree of standard deviation of track geometry irregularity in each lot before and after maintenance.(2) By analyzing the deterioration models of track geometry irregularity,several linear and nonlinear models that can describe typical track irregularity condition are chosen.The track irregularity forecast effects of these two kinds of models are evaluated using inspection data and compared with the real inspection data.The result shows that the track geometry irregularity can be described with linear deterioration model in a certain short length lot.The linear deterioration forecast model of track geometry irregularity for a single lot is es
