全球智库信息集成服务系统

How well do the implemented models predict the mode shares observed in the local household interview data?

• Overall the predicted mode shares are close to the observed values. For some purposes the car driver share is under-predicted which is related to how the model predicts the distribution of persons across different car availability segments.

How well do the implemented models predict tour lengths compared to National Travel Survey data for South East Wales?

• Overall the model validated reasonably well against the tour length information observed in the National Travel Survey, however there was some tendency to over-predict car passenger tour lengths and under-predict train tour lengths.
• Subsequent to this validation step, adjustments were made to the models so that the predicted proportions of tours by distance band exactly matches the National Travel Survey distributions.

This report documents the implementation of the demand models that form part of the South East Wales Transport Model (SEWTM). The model implementation structure and the mode-destination choice model parameters have been transferred from the strategic model developed for the West Midland region, PRISM West Midlands.

The SEWTM travel demand forecasting system comprises three principal components that are either identical or closely related in structure to those used in the PRISM West Midlands model:

• Chapter One

  Introduction

• Chapter Two

  Segmentation

• Chapter Three

  The Population Model

• Chapter Four

  Travel Demand Models

• Chapter Five

  Model validation

• Chapter Six

  Summary and recommendations

• Appendix A

  Segmentations

• Appendix B

  Population Model validation

• Appendix C

  Final processing module