Analog Neural Network for Multivariable Algebraic Loops in Constrained Control

Abstract

An analog neural network is designed to efficiently solve the proposed dynamic multivariable algebraic loop representation of a constrained control problem. The neural network is formulated by introducing dynamics into the static algebraic loop representation of the problem which allows it to be applied to a larger class of control problems. The proposed circuit implementation is practical and can be realized directly with passive components and operational amplifiers or with a field programmable analog array (FPAA). An example with various MATLAB simulations is included to demonstrate the effectiveness of the solution.