GLOBAL OPTIMIZATION IN SEPARATION NETWORK SYNTHESIS

In this paper two global optimization methods are applied to the separ ation network synthesis (SNS) problem. The first is a stochastic metho d of CSENDES [1], the second is an interval method. The latter method used is a two-stage algorithm where in the first stage a close upper b ound for the global optimum is given. The efficiency of these methods is illustrated by solving the SNS problems given by FRIEDLER et al. [2 ] and KOVACS et al.[3].