Trip distribution modelling using neural network
Trip distribution is the second important stage in the 4-step travel demand forecasting. The purpose of the trip distribution forecasting is to estimates the trip linkages or interactions between traffic zones for trip makers. The problem of trip distribution is of non-linear nature and Neural Netwo...
| Main Authors: | , |
|---|---|
| Format: | Conference Paper |
| Published: |
2013
|
| Online Access: | 649 http://hdl.handle.net/20.500.11937/32141 |
| _version_ | 1848753579219746816 |
|---|---|
| author | Rasouli, M. Nikraz, Hamid |
| author_facet | Rasouli, M. Nikraz, Hamid |
| author_sort | Rasouli, M. |
| building | Curtin Institutional Repository |
| collection | Online Access |
| description | Trip distribution is the second important stage in the 4-step travel demand forecasting. The purpose of the trip distribution forecasting is to estimates the trip linkages or interactions between traffic zones for trip makers. The problem of trip distribution is of non-linear nature and Neural Networks (NN) are well suited for addressing the non-linear problems. This fact supports the use of artificial neural networks for trip distribution problem. In this study a new approach based on the Generalised Regression Neural Network (GRNN) has been researched to estimate the distribution of the journey to work trips. The advantage of GRNN models among other feedforward or feedback neural network techniques is the simplicity and practicality of these models. As a case study the model was applied to the journey to work trips in City of Mandurah in WA. Keeping in view the gravity model, the GRNN model structure has been developed. The inputs for the GRNN model are kept same as that of the gravity model. Accordingly the inputs to the GRNN model is in the form of a vector consist of land use data for the origin and destination zones and the corresponding distance between the zones. The previous studies generally used trip generations and attractions as the inputs to the NN model while this study tried to estimate the trip distribution based on the land uses. For the purpose of comparison, gravity model was used as the traditional method of trip distribution. The modelling analysis indicated that the GRNN modelling could provide slightly better results than the Gravity model with higher correlation coefficient and less root mean square error and could be improved if the size of the training data set is increased. |
| first_indexed | 2025-11-14T08:26:45Z |
| format | Conference Paper |
| id | curtin-20.500.11937-32141 |
| institution | Curtin University Malaysia |
| institution_category | Local University |
| last_indexed | 2025-11-14T08:26:45Z |
| publishDate | 2013 |
| recordtype | eprints |
| repository_type | Digital Repository |
| spelling | curtin-20.500.11937-321412018-04-13T06:36:30Z Trip distribution modelling using neural network Rasouli, M. Nikraz, Hamid Trip distribution is the second important stage in the 4-step travel demand forecasting. The purpose of the trip distribution forecasting is to estimates the trip linkages or interactions between traffic zones for trip makers. The problem of trip distribution is of non-linear nature and Neural Networks (NN) are well suited for addressing the non-linear problems. This fact supports the use of artificial neural networks for trip distribution problem. In this study a new approach based on the Generalised Regression Neural Network (GRNN) has been researched to estimate the distribution of the journey to work trips. The advantage of GRNN models among other feedforward or feedback neural network techniques is the simplicity and practicality of these models. As a case study the model was applied to the journey to work trips in City of Mandurah in WA. Keeping in view the gravity model, the GRNN model structure has been developed. The inputs for the GRNN model are kept same as that of the gravity model. Accordingly the inputs to the GRNN model is in the form of a vector consist of land use data for the origin and destination zones and the corresponding distance between the zones. The previous studies generally used trip generations and attractions as the inputs to the NN model while this study tried to estimate the trip distribution based on the land uses. For the purpose of comparison, gravity model was used as the traditional method of trip distribution. The modelling analysis indicated that the GRNN modelling could provide slightly better results than the Gravity model with higher correlation coefficient and less root mean square error and could be improved if the size of the training data set is increased. 2013 Conference Paper http://hdl.handle.net/20.500.11937/32141 649 fulltext |
| spellingShingle | Rasouli, M. Nikraz, Hamid Trip distribution modelling using neural network |
| title | Trip distribution modelling using neural network |
| title_full | Trip distribution modelling using neural network |
| title_fullStr | Trip distribution modelling using neural network |
| title_full_unstemmed | Trip distribution modelling using neural network |
| title_short | Trip distribution modelling using neural network |
| title_sort | trip distribution modelling using neural network |
| url | 649 http://hdl.handle.net/20.500.11937/32141 |