Extending the state-of-the-art for the COMAN/ATCAL methodology

Abstract

This thesis reviews currently existing attrition methodologies and critically examines the (ATtrition CALibration) ATCAL approach for heterogeneous force mixes. Lanchester attrition-rate coefficients must be calibrated for use in ATCAL. When the aggregated combat results from a Lanchester-type attrition model agree with the results of a high-resolution simulation, for a particular mix of weapons on either side, the equation is said to be calibrated. Combat scenario runs are made in the high-resolution JANUS combat model. A maximum likelihood estimation approach (COMAN), which incorporates target priority and availability information, is used to estimate the Lanchester attrition rates in the ATCAL model. A continuous-time, three-state Markov chain model of target acquisition in JANUS is used to obtain the target availability parameters required in the COMAN approach. Bootstrap confidence intervals are developed for the attrition-rate estimates and the target availability parameters. The ability of the ATCAL methodology to replay JANUS results in an aggregated, heterogeneous-force replay model is investigated. These developments have the potential for allowing high-resolution simulation results to be extrapolated to a wider spectrum of conditions (e.g., force mixes) than currently possible by virtue of the more detailed nature of the aggregated-replay model involved.