Integrated optimization of system design and spare parts allocation by means of multiobjective genetic algorithms and Monte Carlo simulation

In this paper, the issue of the combined optimization of system design and spare parts allocation is addressed from a methodological point of view. The problem is framed as a typical optimization problem with multiple choices of component types and number of spares available for each component category. The component types differ with respect to their availability and cost characteristics. The optimization variables are then the type of components (both principal and spare) to be allocated for each category and the number of spare parts to be allocated for each component category in the operating system. The optimization is carried out according to a multiobjective perspective, with the aim of identifying optimal compromising solutions characterized by both high system availability and revenues. The optimization problem is solved by means of an approach that effectively combines genetic algorithms, as the multiobjective search engine of the optimal solution for the system design and spare parts allocation, and Monte Carlo simulation, as the evaluation engine of the system solution performance with respect to the availability and revenue objectives (the fitness functions of the genetic search algorithm). The approach has been previously introduced in works coauthored by the first author of this paper and demonstrated to achieve statistically accurate estimates of the fitness functions for the potentially optimal solutions, without wasting computational resources on solutions with low potential for optimality. Two numerical examples are presented.