الفهرس 
Introuduction…Only 14 pages are availabe for public view
 |  | from | 135 |  |  |
| | | from | 135 | | |
Abstract
Selection of an appropriate Maintenance and Rehabilitation (M&R) treatment for flexible pavements is often a complex process. Some researchers have suggested a “decision tree” approach to select the most feasible repair strategy based on the existing pavement condition. A simplified innovative decision tree system, Maintenance Unit (MU), was developed through a research conducted at Cairo University, Egypt. The MU system determines M&R activities based on the density of distress repair methods (not the density of individual distresses). In addition, it addresses the complex combination between distress levels and maintenance alternatives. Predicting maintenance decisions is also an important component for a multi-year analysis, which is considered a complex procedure. A maintenance decision model (MDM) using the MU system, was developed to predict the future maintenance decisions through a previous research published in International Journal of Pavement Engineering. Due to the complexity of such a decision system, automated implementation is vital and needed.
Artificial Neural Network (ANN) approach can be used for the elimination of this drawback. This research aims at employing the inherent capabilities of ANN technology in determining current and future pavement maintenance decision. This necessitates to create inclusive database through data collection procedures as well as creating design cases.
The General Authority for Roads & Bridges and Land Transport (GARBLT), Ministry of Transport, divides the Egyptian main roads network into twelve (12) districts. Administration of District number 6 is located in Beni Sueif governorate and it is responsible for all major roads surrounding Beni Sueif area including Fayoum governorate. There are 17 main roads in district #6 and the total length of these roads is around 698 km. In this study, data collection was conducted on the GARBLT main roads in Fayoum region. Four roads have been selected for this study with total length of 51.3 Km. The selected roads are as follows:
• Fayoum Ring Road, 20 km length
• Fayoum - Itsa Road, 7.2 km length
• Fayoum - Abshaway Road, 14.4 km length
• Senores - Tamia Road, 9.7 km length
Consequently, a database was created including all existing distresses (problems) that could be found on the asphalt surface of these selected roads. Current and future pavement maintenance decisions using MU system and MDM model, respectively, have been identified. This creates two sets of databases for current and future maintenance decisions, which were used to train and to test the developed neural network using MATLAB program, version 7.12.
Two ANN-based maintenance decision models have been developed to recommend current and future pavement maintenance decisions based on MU system and MDM model, respectively. Results of this study reveal that artificial neural network is appropriate for implementation in recommending current and future flexible pavement maintenance decision. This is particularly promising for developing countries where such applications can play an effective role in offsetting the lack of decision tools, which is often apparent.