Providing a Model Based Fuzzy-Neural Network Logic to Predict the Amount of Trip in the Medium Cities (Case Study: Yazd, Iran)

Analysis of trip demand has four stages and trip generation is one of the initial and most important stages aims to estimate number of generated trip in different regions of the city accurately. Since, the obtained information of trip generation stage is one of the main data and influences on accuracy of all stages in analysis of other trip stages, so increasing prediction accuracy in trip generation stage can provide more accurate results of model. Increasing prediction accuracy is possible in two ways generally. First, increasing accuracy and dimensions of used database which is expensive (in term of financial and time), and second is using more advanced modeling methods. So, in this study in addition to identify the influencing factors on trip generation and providing some descriptions about fuzzy systems and how to train these system due to information limit, a hybrid model for trip generation based on linear regression and fuzzy- neural system is provided as an appropriate option. For this purpose, eight important parameters including population, number of students, number of college’s student, per capita ownership of cars in household, distance of traffic area of the city center, number of employed individuals in the region, number of residents employed individuals and number of resident households are used as most important influencing factors in the trip generation as input function and number of generated trips is used as output function. Finally, this modeling method is used to process models of Yazd city based on transportation information in 2012 and the results comparison of high accuracy of model, despite additional cost is needed to supply information.

Providing a Model Based Fuzzy-Neural Network Logic to Predict the Amount of Trip in the Medium Cities (Case Study: Yazd, Iran) Keywords:

Trip Generation , Fuzzy Set , Blend Model , Fuzzy-Neural Networks , Engineering Estimation