Application of neural networks to sorting problems

The authors present a design philosophy for constraint optimization problems which is based on viewing the neural network as an interconnection of several subsystems. Specifically, they present a method to classify ranges in the magnitude of an analog input signal and associate an arbitrary digital output with each range. The proposed approach is illustrated by application to a cascadable 2-bit analog/digital converter. The advantages of the proposed design method as compared with energy function designs are enumerated.<
>