Developing Trip Generation Models Using Adaptive Neuro-Fuzzy Inference System: Salfit City as a Case Study
| dc.contributor.author | Irshaid, Mohammad | |
| dc.date.accessioned | 2022-10-10T06:50:06Z | |
| dc.date.available | 2022-10-10T06:50:06Z | |
| dc.date.issued | 2019-12-23 | |
| dc.description.abstract | In Palestine, few studies that are concerned with the development of trip generation models have been conducted. The lack of specialized studies for this purpose may be related to several challenges that encounter the Palestinian situation, such as the restricted financial support and the lack of reliable data, which makes it difficult to perform such studies. These limited studies were developed using mainly the Multiple Linear Regression (MLR) approach, which sometimes would not result in appropriate models when dealing with interrelated and complex relationships among several socioeconomic variables. Therefore, this study was devoted to investigating the feasibility of using a relatively new method for data analysis called the Adaptive Neuro-Fuzzy Inference System (ANFIS), as an alternative for the traditional MLR, and explore its application within the Palestinian context for the development of the home-based trip generation models. Through this study, four types of trip generation models were developed for the Palestinian city of Salfit; the ALLTRIP model for estimating the total number of daily home-based trips generated, and the other three models for estimating the number of trips generated based on trip purpose, which are the Home-Based Work (HBW), the Home-Based Education (HBE), and the Home-Based Other (HBO) trips generation models. These models were estimated and validated using a sample of 309 households, that was thoroughly collected for Salfit City in 2017. Each of these models was developed using the two competing approaches; MLR and ANFIS. The better performing and more suitable approach was then determined based on several evaluation criteria, such as the higher value of R-Squared, the lower RMSE, and the much closer outputs to the actual values. In this study, the ANFIS was able to outperform and develop more accurate models than the MLR when dealing with the ALLTRIP and the HBO, which were considered to be more complex than others, as they include wider data range, and constitute more percentage of daily trips generated. Whereas for the HBW and the HBE, both modeling approaches were performed nearly at the same level, the R-Squared values were large enough to capture most of the variations, and the differences between the performance measures were very small which could be neglected. On the other hand, there was a little advantage for the MLR in the validation process. For these two cases, the use of the MLR was considered to be sufficient. The robust comparison through this study reveals that the ANFIS represents a promising technique, that could be a good competitor for MLR approach, especially, when dealing with interrelated and complex relationships among several socioeconomic variables. The ANFIS was found to be a useful tool for modeling home-based trip generation for Salfit City, and its further applications in transportation planning studies were recommended. | en_US |
| dc.identifier.uri | https://hdl.handle.net/20.500.11888/18030 | |
| dc.publisher | An-Najah National University | en_US |
| dc.subject | Developing Trip Generation Models Using Adaptive Neuro-Fuzzy Inference System: Salfit City as a Case Study | en_US |
| dc.supervisor | Prof. Sameer Abu-Eisheh | en_US |
| dc.title | Developing Trip Generation Models Using Adaptive Neuro-Fuzzy Inference System: Salfit City as a Case Study | en_US |
| dc.type | Thesis | en_US |