NEURAL-NETWORK OPTIMIZATION FOR REDUNDANCY ALLOCATION

System reliability optimization problems such as redundancy allocation are hard to solve exactly. Neural networks offer an alternative compu tational model for obtaining good approximate solutions for such probl ems. In this paper we present a neural network for solving the redunda ncy allocation problem for a n-stage parallel redundant system with se parable objective function and constraints. The problem is formulated as a 0-1 integer programming problem and solved using the network. The performance of the network compare favourably with that of the best f it algorithm. The number of iterations taken by the network increases very slowly with increase in number of variables. Hence the network ca n easily salve large problems.