The effect of uncertainty on Lanchester type equations of combat

Abstract

This thesis examines whether the complex random process of combat can be adequately represented by a deterministic model. Does one destroy any of the essential features of the random combat process by considering a deterministic model as representing the "average" course of combat? Insights into the fundamental differences between deterministic and stochastic models are obtained by comparing the deterministic and stochastic version of the so-called Lanchester "square-law" attrition process. three aspects of the models are compared, with several hypotheses examined for each: possibility of winning, the expected force level time history, and the variance of the expected force levels. From the analysis it is concluded that if the forces are not near parity, and if the initial force levels are relatively "large", a deterministic model can adequately represent combat.