A queuing model for designing an optimal three-dimensional maintenance float system

In this paper, we consider a float system that consists of a single workstation, a repair station of several parallel and identical servers, and an inventory buffer of several standby components serving the workstation. We determine the optimal capacity of inventory buffer (F), the size of repair crew (S), and the mean repair rate (μ) such that the total operations cost is minimized. In particular, by assuming exponential failure and repair, we model the system as a finite birth and death process (an M/M/S/F queue). We develop a generalized optimization algorithm to obtain the optimal combinations of all three variables.