A Quantitative Approach to Aggregation in the Modeling of Tactical Command, Control, and Communication Systems

摘要

This research effort is directed at determining the feasibility of using structural-sensitivity measures as the basis for simplifying large mathematical models of dynamic systems (e.g., models of tactical command, control, and communication systems). Toward this end, a FORTRAN program was written which can be used to simplify models characterized by sets of linear differential equations. Specifically, the program determines the optimal simplified model (i.e., the set of coefficients characterizing the set of linear differential equations) of specified dimension corresponding to the original linear model of higher dimension. The algorithm, on which the FORTRAN program is based, minimizes an objective function which is defined in terms of the structural sensitivities of the state variables to be preserved in the simplified model. The program inputs are the set of coefficients which define the linear differential equations representing the original model, the dimension of the simplified model to be determined, and the set of parameters defining the objective function to be minimized. The program outputs are the set of coefficients which define the optimal simplified model and the value of the objective function corresponding to the optimal simplified model. (Author)