An analysis to assess the value of information when freight origin-destination (O-D) tables are estimated is developed. The theory of design of experiments is used as a framework to ascertain the value of three categories of information: link counts, O-D totals, and aggregate O-D totals. On the basis of these insights, a forecasting model is developed for estimating the table quality to be obtained given the type and quantity of information used. Finally, the effect of introducing errors into the information is explored.
Abstract