IMPROVED METHODS FOR TRIP GENERATION

The acceptance by the UK Government of the SACTRA recommendation that induced traffic should be taken into account in road scheme assessment has focused modelling interest again on the methods by which the various components of induced traffic can be forecast. The possibility that increases in the number of trips may be among the relevant changes means that appropriate models of trip generation need to be considered. Drawing on recent research for the European Commission and the Department of Transport, as well as on other work, the paper addresses the problems of formulating and estimating appropriate models of trip generation, considering a number of the issues that arise: a) the connection between trip generation and the activity basis of travel demand; b) the contrasts between household and person-based models and between trip and tour modelling; c) the nature of the data that is available for modelling; d) the impact of the inclusion of slow modes in the trip rates; e) the ways in which accessibility can be measured and the incorporation of generation models in hierarchical logit model systems; f) the appropriate form of model to be used and its relationship to demand elasticity; and g) the distinction between generation and attraction modelling. It is shown that quite simple models can be developed which allow an efficient use of data to test the correlation between trip rate and accessibility and in particular that very simple methods can be used in 'pivot point' procedures to predict changes in trip rates resulting from changes in the 'level of service' offered by transport networks. Practical results are given from a number of studies to illustrate the sensitivity of trip generation to accessibility in various contexts. For the covering abstract, see IRRD E101013.