A model is presented for developing the state probability distribution
of aggregate space requirements in a randomized storage system as a f
unction of the level of variation in item retrieval activity. This mod
el provides a tool to analyse tradeoffs in space requirements and retr
ieval efficiency associated with dedicated and randomized storage poli
cies. The technique extends previous work by accounting for space redu
ctions associated with randomized storage in estimating average item r
etrieval costs. A sample problem is presented along with an analysis o
f randomized and dedicated storage policy performance.