A Soluble Theoretical Model of the Probable Amount of Material in an Inventory

Anybody at all familiar with the operation of a nuclear fuel fabrication plant knows that it is quite presumptious to set up a mathematical model to represent the complex work- ings of that plant, even for such a limited purpose as a safeguard's study. Yet, it is the very fact of that com- plexity which makes attractive some sort of an operations research approach. In discussions concerning setting up some computer-based operations research calculations, it occurred to us that many of the probabilistically determined processes which one would have to mock up on the computer for an operations research simulation study, are amenable to treatment by what is called in probability theory a Markov chain. To the extent this is the case, it then followed that it would be possible to set up analytic models of processes which would otherwise have to be simulated by Monte Carlo techniques. Even if the proc- ess had to be drastically simplified to fit the format of a Markov chain, it might still be worthwhile to do so, in that the analytical Markov calculation would furnish a standard with which to compare the more accurate, but much more com- plex and troublesome, Monte Carlo calculation.