الفهرس 
ContentsOnly 14 pages are availabe for public view
 |  | from | 120 |  |  |
| | | from | 120 | | |
Abstract
The use of technology in education is transforming teaching and learning into a different educational experience. The digital electronics revolution keeps us connected with almost anyone around the world and makes information available any-where, at any-time. According to eMarketer Statistics, by this years’ end nearly 70% of college mobile phone users will own smartphones. In this thesis, we will: (i) investigate the benefits and challenges of current M-Learning systems; (ii) try to overcome some of the M-Learning technical challenges such as battery life, wireless network availability, and shared network bandwidth; (iii) develop a control mechanism for optimizing data transfer to mobile devices based on a proposed hybrid online/offline model that minimizes the power consumption of mobile battery and decreasing on network traffic for mobile devices; (iv) try to enhance the battery power consumption of mobile device by designing a proposed Schedule/Cache based architecture and for delivering and caching the contents efficiently to mobile users, and (v) evaluate the proposed model in terms of scheduling algorithm, and caching policy using battery power consumption of mobile device compare them with the traditional and current schemes to verify the improvement of the new model and using performance metrics for caching techniques to the best one. For proposed and implementing hybrid online/offline M-Learning model based architecture, using knowledge of the relevance and current network/system conditions. In the adopted simulation scenarios, by using Python 2.7.12 programming we have created 500 files with random size between 500KB and 5GB on the server side and made 10,000 random requests with Gaussian random for file popularity. The simulation results showed that, the proposed model improves: (i) Battery Life Time: mobile learning device battery life up to 28.6 % based on Wi-Fi communication technology, 66.7% for 3G connections comparing without using proposed model, and using the same size of data transfer, and the proposed model with Wi-Fi connection can saved battery life up 14.3 % over using 3G connection and (ii) By using Matlab 8.2R 2013b , Java SE8, and cache simulator (CS) for caching Algorithm: Least Frequently Used-Server Statistics (LFU-SS) algorithm is the best caching algorithm for small cache sizes (8MB- 64MB) which achieve up to 11% of improvement over Least Recently Used (LRU) algorithm cache policy, achieved 32% of saved byte ratio, and data transfer decrease ratio up to 68%. On the other hand, for large cache size (128MB-512MB) the Least Recently Used-K (LRU-K) caching algorithm introduced the best performance, which achieved up to 10% of improvement over LRU cache policy, it saved up 92% of the network traffic and data transfer decrease ratio up to 8.7%. Based on these extensive simulation results, we have proposed a modified caching algorithm to provide optimal performance for all caching sizes. These results were verified using both Performance matrices and ROC curves.