Improvements in calibration and understanding of the competing opportunities model of trip distribution
This note suggests a method for improved calibration of the competing opportunities model of trip distribution. The method, called COMNEST, includes three steps: (1) selection of a suitable uniform width of time bands for the study area; (2) determination of the width of the first time band to achieve best-possible results; and (3) improvement in the accuracy of prediction of the dependent variable for the zones within the first time band. The third step involves a redistribution of the total number of trips predicted for the first time band among the traffic zones within the first time band by reapplying the competing opportunities model concept. Using the examples of trip distribution for Edmonton, Winnipeg, and Washington, DC, it is clearly shown in this note that the application of the suggested method can considerably increase the accuracy of the competing opportunities model. Key words: trip distribution, competing opportunities model, urban transportation planning, traffic forecasting, urban planning.