This thesis explores the validation of Lanchester equations as models of the attrition process for the Battle of Kursk in World War II. The methodology and results of this study extend previous validation efforts undertaken since the development of the Ardennes Campaign Simulation Data Base (ACSDB) in 1989 and the Kursk Data Base (KDB) in 1996. The KDB is a computerized database developed by the Dupuy Institute and the Center for Army Analysis from military archives in Germany and Russia. The data are two-sided, time-phased (daily), highly detailed, and encompass 15 days of the campaign. The primary areas of analysis are the effect of using purely engaged forces in parameter estimation and the effect of force weighting in forming homogeneous force strengths. Based on the numbers of personnel, tanks, armored personnel carriers, and artillery, three different data sets were constructed: all combat forces in the campaign, combat forces within contact that are both engaged and not engaged, and combat forces within contact that are engaged. In addition, a weight optimization program using a steepest ascent algorithm was developed and utilized. Findings indicate that Lanchester-based models provide a considerably better fit for data sets composed only of forces that are actively engaged. Also, Lanchesterâ s linear model appears to provide the best fit to the Battle of Kursk data. Finally, optimization of force weights does not significantly improve the fit of Lanchester models.
展开▼
