Urban Trip Generation using Fuzzy Logic Based on a Case Study in the City of Rasht

Updating and predicting different types of Trip Generation Models (TGM) is always known as a challenge in different studies. Trip Generation (TG) is the first and one of the most important stages between fourfold trip demand prediction stages that its goal is to estimate the total number of generated trips from one source. Researchers such as Ortuzar and Willumsen emphasized that more non-congestion trip generation models have better prediction and adaptation ability. So in many years to validate different ideas, experts have invented variety of models and they always try to estimate the number of generated trips in a better method helping newer model presentation. Also they try to decrease the previous models problems and defects as much as possible. At this research a new method using fuzzy logic was presented to predict the number of generated trips using trip generation data in City of Rasht, in north Iran. For configuration of the fuzzy model four important parameters have to be considered: household structure, family size, car ownership and income. Data of these parameters are integer and converting the integer numbers to fuzzy form is meaningless. Therefore, three parameters; household structure, family size and car ownership were normalized to income that is one of the most important parameters at trip generation. Then to develop fuzzy model, the ratio of household structure, family size and car ownership to income was used as input functions and the ratio of number of generated trips to income was used as output function. The basic idea at this method is using fuzzification-defuzzification procedure and selecting input-output function based one data of one research in City of Rasht. One of the more common methods to develop TGMs is step-wise regression method. To represent very high accuracy of proposed fuzzy method, this model was considered and compared with step-wise regression analysis, step by step. Because of three parameters: family size, car ownership and income were identified as the most effective parameters in TG, at first fuzzy model was developed versus these three parameters, then proposed model was developed using four parameters: household structure, family size, car ownership and income. To compare with step-wise regression method, 4-variable step-wise regression equation was used. As we predicted before, the results of 4-variable and 3-variable step-wise regression equations do not have any differences, because these correlation coefficients have only 0.1% difference. The result of fuzzy model with 4 parameters and 3 parameters represent that these two models have differences remarkable and using household structure-parameter as an extra input leads to exact accurate results. To compare, these two methods, fuzzy model and step-wise regression analysis, were performed on 20 groups of random statistical data of City of Rasht that was selected with MATLAB software. Result of these comparision shows that in step-wise regression method except one that can originate from data dissipation maximum error was 22%, but fuzzy method approximately without error leads to very accurate results. Fuzzy method error was zero for all data except one. In addition to fuzzy method leads to correct and very accurate results for all ranges of data, but step-wise regression method results correctly only for given ranges of data. Another benefit of fuzzy method is that this has less mathematical complexity with the compare of step-wise regression method. According to above-mentioned consideration, it can be concluded that proposed fuzzy method at this research has high accuracy compared to other methods and this model is one of the most assured methods for prediction of generated trips.