PMAKE Analysis: Predicting Rail Yard Time Distributions Using Probabilistic Train Connection Standards

Abstract

Any study of rail freight service or freight car utilization must eventually address the reliability of train connection performance. In a series of case studies over the past 10 years, the MIT Rail Group developed PMAKE analysis as an effective means of doing this. The PMAKE methodology has been used extensively in service planning by a number of railroads and has been used in two major rail network models. The Freight Car Utilization Program has supported two demonstration projects showing how to incorporate PMAKE analysis in railroad terminal control systems. The paper reviews three other major modeling approaches used in rail yard studies: simulation, queuing, and capacity planning. These approaches have had limited effectiveness in prediction train connection performance and in supporting management control systems. PMAKE analysis improves upon these approaches by making it easier to (a) calibrate models of yard performance and (b) incorporate analytic techniques within the operating control system. In short, PMAKE analysis builds a bridge between operations research technique and management control. The paper includes the parameters of PMAKE functions calibrated for various rail yards. It also includes the mathematical basis for PMAKE analysis and shows several means of calibrating PMAKE functions. This paper, therefore, provides a complete description of PMAKE analysis, including its development, its relation to other methods of terminal analysis and its use by the rail industry.