Biased model reduction of discrete interval system by differentiation technique

A number of techniques for the reduction of interval system have been presented by various researchers. But the validity of the method is based on the resulting error by the model reduction. The system with parameter variations within bounds creates intervals in the coefficients of the system polynomial; hence the system is called interval system.In this paper, a differentiation technique is applied to interval system to obtain, lower order stable models retaining the initial Markov parameters and time moments of the original system. These models give better approximation, for both, steady state and transient characteristics of the time response. Application of Routh type arrays, reciprocal transformations and the time moments of the nth order original system are all avoided in this model reduction of interval system, otherwise it will further complicate the analysis of interval system. So the new technique applied to the interval system, is a direct method, simple and computationally better than the other methods, involving differentiation technique.