Logic response statistics are developed for a class of randomly interconnected zero memory logic nets which are frequently used in the so-called learning systems. From these statistics, it is shown that it is possible to obtain the probabilities of the transmission characteristics of the randomly interconnected logic networks once the interconnections probabilities are known. The method of identification shown here follows one developed in an earlier paper of Wolf and Wolf and Diets with minor modifications to account for the logic nature of the networks. Following the previous wo~.h, an ergodic probe is used as the stimulus to excite all modes of operation of the zero memory logic networks in order to identify the transmission probabilities. The probe in this case is not white noise, but rather binary noise.
An example is shonm of a logic system consisting of two inputs and a single output with randomly interconnected elements. This logic system is identified in terms of the probabilities of interconnection and the logic response statistics. lntroductlon 1
In the literature, Wolf (1) and Wolf and Dietz (2) discuss general techniques for probing continuous analogue systems with stochastic processes to charaeterize and identify system and physical parameters. The principle of probing * The opinions or assertions made herein ~re those of the authors and are not ~o be construed as official or reflecting the views of the Navy Depart~ment or the Naval service at large.