PREDICTING URBAN TRIP GENERATION USING A FUZZY EXPERT SYSTEM

One of the most important stages in the urban transportation planning procedure is predicting the rate of trips generated by each trafic zone. Currently, multiple linear regression models are frequently used as a prediction tool. This method predicts the number of trips produced from, or attracted to each trafic zone according to the values of independent variables for that zone. One of the main limitations of this method is its huge dependency on the exact prediction of independent variables in future (horizon of the plan). The other limitation is its many assumptions, which raise challenging questions of its application. The current paper attempts to use fuzzy logic and its capabilities to estimate the trip generation of urban zones. A fuzzy expert system is introduced, which is able to make suitable predictions using uncertain and inexact data. Results of the study on the data for Mashhad (Lon: 59.37 E, Lat: 36.19 N) show that this method can be a good competitor for multiple linear regression method, specially, when there is no exact data for independent variables. 1. Introduction The first and one of the most important steps in the process of trafic volume prediction is to estimate the number of trips generated in each trac zone. Two widely-spread methods that are used for prediction are \multiple linear regression" and \cross classification" techniques [21, 32, 37]. A cross-classication table shows the number of daily trips produced per house- hold for dierent characteristics of households, such as household size, number of owned cars, etc. The trip rates per household are developed from a household survey, and will be used for predicting the number of future trips. The `cross-classification analysisʹ method is based on estimating the rates of trip production as a function of dierent socio-economic characteristics of households, such as household size, number of owned cars, etc.. The basic assumption is that trip rates remain constant over time for aforesaid stratication. Therefore, the number of households of each household category needs to be forecasted in the future [37]. Multiple linear regression (MLR) method, relates the number of trips (or trip rates) of each purpose generated by each individual, household, or trafic zone, Received: January 2011; Revised: September 2011; Accepted: October 2011 Key words and phrases: Trip generation, Multiple linear regression, Membership function, Fuzzy rules, Fuzzy expert system.