A Comparison of a Location Level Simulation Model for Multiple Service Demand Forecasting with Other Models
25 January 1988
A Monte Carlo simulation approach to multiple services demand forecasting can be shown to reproduce the mean demand for services as calculated from the logit or probit model. Some likely bounds on the accuracy of the estimates based on the simulation model are derived and used to estimate the number of simulations needed to achieve the desired accuracies. Examples are given to illustrate the comparisons of the mean demands calculated from the simulation, probit and logit models. It is further discussed that at the location level, the simulation approach is potentially more flexible, realistic and easier to use than the commonly used logit or probit model.