Bayesian design in queues: An application to aeronautic maintenance

We exploit Bayesian criteria for designing an M/M/c//r queueing system with spares. This problem arises in many applications of the so-called machine interference problem, like job-shop type systems, telecommunication traffic, semiconductor manufacturing and transport. We exemplify our approach by a problem from aeronautic maintenance, where the numbers of repair crews c and spare planes s must be sufficiently large to meet the necessary operational capacity. Bayesian guarantees for this to happen can be given using predictive or posterior distributions. The differences between the several arising possibilities are illustrated in the paper.