A neural network for Euclidean distance minimization

Abstract
An electronic neural network for the Euclidean distance minimization problem, implemented in VLSI-based hardware, is described. The convergence properties of the neural-network hardware are investigated and compared with computer simulation results. The neural network's ability to find the 'best' or a 'good' solution is quickly demonstrated. The effect of noise on the circuit dynamics and the convergence behavior of the neural network hardware is also examined.