TESTING THE PRACTICAL VALUE OF THE UMOT MODEL

The UMOT model was developed by Yacov Zahavi as an alternative to conventional transport demand models. Zahavi claimed that UMOT was very economical, that was based on sound behavioural assumptions and observed regularities. It was further claimed that the model was policy sensitive, transferable over time and space and required a minimum of adaptation to each study area. These latter claims have been tested from a pragmatic point of view utilising data from Washington in the USA and Belgrade and Osijek in Yugoslavia. Our findings did not support Zahavi's claims. Both two-mode and three-mode versions of UMOT required tuning to reproduce survey data and in some cases either model simply failed to converge. Even the policy sensitivity of UMOT was in question when it was found that under certain circumstances a reduction in transit fares generated more travel by car. Some possible uses for UMOT and improvements are discussed at the end of the paper.