This paper presents some innovations for overcoming shortcomings in the current state-of-the-art for the hierarchy-of-models approach to modeling aggregatedforce attrition in ground-combat models. The basic concept of such an approach for modeling large-scale system behavior is presented, together with the theoretical underpinnings for modeling attrition in large-scale ground combat. The output of an entity-level discrete-event combat simulation is fit to a Lanchester-type aggregatedreplay model. Use of a reliable statistical-estimation technique for determining model parameters is emphasized. The main innovation is to show how use of more detailed output data (e.g. line-of-sight (LOS) data) from the high-resolution simulation allows one to develop maximum-likelihood estimates. The methodology is applied to a current high-resolution DoD combat model, and a Lanchester-type aggregated-force replay model is developed.
James G. Taylor, Ugur Ziya Yildirim, William S. Mu