Scalable GEH: A Quality Measure for Comparing Observed and Modeled Single Values in a Travel Demand Model Validation

Abstract

For proving the validity of a travel demand model, the deviations between observed values and modeled values must be quantified and evaluated by suitable quality measures. The comparison of observed and modeled values can refer to single values, sets of single values, and distributions of values. The paper focuses on comparing pairs of single values with specific reference to the quality measure GEH (named after Geoffrey E. Havers, who introduced it for traffic planning purposes in the 1970s), which is used in the British guidelines WebTAG. The paper describes desirable and problematic properties of the GEH and examines how it is related to normal distribution. With the objective of overcoming the limitations of the GEH an alternative quality measure was developed, which can be scaled to validate different value ranges representing hourly or daily traffic volumes, daily number of trips per person, or trip distances per trip purpose. The paper concludes with a proposal for validating single values.