Optimal design problems of highly reliable systems and their solution methods by surrogate constraints algorithm

This paper presents a solution method for the reliability optimization problem of systems which is formulated as a mixed-integer nonlinear programming problem with multiple constraints. The system structure is series, series-parallel, or complex. In the solution method, the optimal solution is obtained by solving the surrogate dual problem. Surrogate problems with only one constraint which appear in the optimization process are solved by a technique using dynamic programming. Some computational experiences are shown along with comparison with other approaches.