To estimate the number of trips generated from a traffic analysis zone or household is the first step of the process of predicting future traffic volume in a region.  A trip generation model expresses the relationship between the socioeconomic factors of the population or the household in an area and the number of trips generated from the area.  Traditionally, this model has been formulated in linear regression form based on the notion that a deterministic relationship exists between the cause and effect.  The deviation of the estimated volume from the actual volume, hence, is considered to be a random error.  This study introduces the notion that the cause-effect relationship of trip generation is perhaps only probabilistic and proposes the use of a fuzzy inference approach for estimating trip generation.

        The fuzzy inference technique has been applied successfully to a number of problems in which causal relationships are considered inherently approximate.  It is a rule-based representation of causality most often found in human and societal behavioral patterns.  In the mathematical formulation of the inference, the approximate nature is accounted for by the use of fuzzy sets.

        This study develops two trip generation models both using the fuzzy inference technique: for a traffic analysis zone based model and for a household based model.  The former estimates the total trip generation in a traffic analysis zone given the socioeconomic characteristics of the zone.  The latter estimates trip generation of an individual household given household characteristics.

        This report:  (1) presents a review of the differences between the traditional regression analysis approach and the fuzzy inference approach; (2) develops the structure of the fuzzy inference model; and (3) compares the estimated trips by the model with the actual numbers and also with the results from DelDOT's current models.