Abstract
This paper presents a new approach to maximize the steady-state availability of a closed queuing multiechelon repairable system with cold standbys. The new method is based on the use of the Taguchi inner array design in simulation modeling of this problem. Regression models are developed based on the results generated from the balanced simulation design. Two special cases of the exponential and Erlang-2 service distributions are presented. These models were validated and shown to provide good approximations to simulation results. Within the specified domain of the problem's parameters, the total operating cost of this system is minimized to determine the optimum spare and repair units.