In this research, a safety management system is planned for a mine operation. The objective is to develop a mine production design that will achieve the desired reliability and variance of reliability estimation of system while performing all sub-system functions at a minimum cost. The problem was formulated in an underground mine system containing five sub-systems with varying reliability estimates, estimation variances and feasibilities and solved by genetic algorithms (GA). (c) 2005 Elsevier Ltd. All rights reserved.