Activity sampling in a stochastic environment

Abstract An alternative model to the one implicitly assumed in conventional activity sampling is offered. Time is said to be partitioned into a series of disjoint intervals. The proportions of time spent on the activity of interest, in these time intervals, are assumed to be i.i.d. random variables. Procedures are developed for estimating moments and other parameters, of the distribution of these variables, from a series of observations. The number of observations required for the proposed approach is demonstrated and discussed.