Controlling algorithmic blocking: Calibration and Cost Redistribution approach to real-world operational railroad application

While algorithmic blocking has been shown to have significant advantages over a traditional, tag table approach for the operational implementation of a freight railroad's blocking plan, railroads have been reluctant to pursue that approach, partly due to the perception that algorithmic blocking is difficult to control. Control refers to configuring the data underlying algorithmic blocking so that block paths returned for rail car movements are those desired by railroad management. This paper presents three complementary control strategies-Every Day, Calibration and Block Cost Redistribution-which together provide a practical way to achieve control with algorithmic blocking. The paper also includes a comprehensive description of an implementation of algorithmic blocking upon which these control strategies are based.