An Alternative to Monte Carlo Sampling in Stochastic Models

The computer resource requirements of Monte Carlo sampling are a serious deterrent to its use in some instances. In such applications as large-scale planning systems which must combine forecast information from many products, processes, and markets, the extensive use of random sampling can become the principal roadblock to further expansion, especially if interactive timesharing applications are indicated. In this paper, we present an alternative scheme for approximating the probability distribution of sums, differences, products, and quotients of independent random variables whose 10th, 50th, and 90th percentiles are specified in advance.