Davis, Gary A.2013-08-282013-08-281994-05https://hdl.handle.net/11299/156001Any proposed traffic management action is essentially a forecast that the action will result in certain traffic conditions, but uncertainty concerning the amount and distribution of traffic demand will introduce random error between what is expected and what actually occurs. This report treats the problem of forecasting whether or not a given set of freeway on-ramp volumes are likely to cause over-capacity demand at some point in the freeway mainline. The main source of uncertainty in these forecasts concerns the freeway's origin-destination matrix, and four different methods for estimating this matrix from loop detector data are evaluated using Monte Carlo simulation. Only the method which explicitly modeled freeway traffic flow produced reasonably unbiased and efficient estimates, and it was concluded that successful estimation must be coupled with a good model of freeway traffic flow.en-USTraffic flowFreewaysOrigin-destination matrixMarkov processesRamp meteringReport