The determination of digital simulation models for continuous systems by direct-search minimization

• Authors: G. G. Leininger;H. R. Martens
• Venue: ACM '68 Proceedings of the 1968 23rd ACM national conference
• Year: 1968

The representation of system characteristics by a discrete time model poses a fundamental problem in the field of digital simulation. In the majority of circumstances the problem consists of approximating a continuous time model well-defined in terms of a transfer function or state variable equation by difference equations or other forms convenient for computer processing. Examples representative of this approach are the Tustin method,1 Boxer-Thaler method,2 the Madwed-Truxal method,3 or the State Transition method.4 Equally important in the simulation field are the numerical integration methods such as the RungeKutta-Blum technique,5 the Runge-Kutta-Merson technique,6 or the Adams-Moulton technique.7 The latter class of techniques is applicable when the system equations are given in terms of first-order (linear or nonlinear) differential equations. A summary and comparison of the above techniques is given by Fryer and Schultz8 and Martens.9 The derivation of discrete analogs for continuous systems has also been discussed by Shaw.10