Optimal availability design of a maintained system

Abstract

The availability of a maintained system depends very much on the number of spares, redundancies and on the number of available repair facilities allocated to individual subsystems constituting the system. The problem formulated in this paper takes into account all these three factors by employing them as decision variables. Since all these three decision variables can only assume integer values, the availability optimization would necessitate a non-linear programming formulation with integrity conditions on decision variables. There are not many elegant and economical algorithms (with respect to time and memory requirement) available in the literature, if an exact integer solution is sought. The purpose of the present paper is, therefore, to provide a complete formulation of availability optimization with respect to spares, redundancies and repair facilities, besides providing a powerful selective search technique, which basically requires few functional evaluations and offers an exact solution. The algorithm is very fast and efficient. Various examples on system design are also provided, which substantiate the necessity of employing all these three decision variables in the formulation if a global optimum is intended for availability maximization problem.