Upton S.C. A statistical mechanics model of combat

Naval Postgraduate School. — Monterey, California, 1987. — 87 p.The purpose of this thesis is to model a Combat System utilizing modern methods of nonlinear nonequilibrium statistical mechanics. This initiates development of methods which eventually can be used as a decision aid to the commander in force planning, battle management, budgeting decisions, doctrinal evaluations, and combat analysis. A general method is developed and then applied to a particular battle scenario using the combat wargame JANUS. The method fits empirical data to a functional form (a Lagrangian) which describes the short time probability distribution of a set of order parameters. A maximum likelihood fit is obtained simulated annealing optimization algorithm. The most likely states of the order parameters and the associated risks (variances) of those states ultimately depend on the detailed structure of the Lagrangian. A long time probability distribution of the order parameters can then be found by using the path integral.