以機率和圖論分析行動學習之學習模型

Abstract

摘要 無線網路及行動裝置已普遍商業化，這些硬體使得數位學習一日千里。像video-phone 和Multimedia-on-demand這些應用也使得多體在這個領域更有發展。 目前還有許多在行動學習上的研究像 Location Management、QoS等，在此環境中資料分散在不同處，我們的研究是找出一個放出一個放置資料的方來提升存取的成功率，同時使用moving pattern來改善資料放置的效能。假設學生要存取不同課程，我們會根據目前的網路狀態定一個成功的機率，同時安排課程也是一個議題，如何在有效的時間內來學習是一項挑戰。 這在這篇論文裡，我們用moving pattern和data allocation來改善在行動學習環境下的存取效能。在課程安排的方法裡會提供學習路徑以及如何在時間內學完，在我們提出的方法中會用實體和邏輯上兩種角度來看，實體面主要探討存取成功率，而邏輯面探討課程安排。 最後我們會一些例子來看這方法的效能如何，也會在不同的網路型態下來研究應用

Abstract Wireless networks, handheld and wearable devices, and devices to sense and control appliances are now a commercial reality. Significant hardware progress in recent years has stimulated the ubiquitous e-learning environment. Applications such as video phones, multimedia-on-demand, and so forth will create new opportunities to access multimedia-rich information in mobile learning environments. There are many areas currently under investigation in mobile learning environments, such as location management, quality of service, and bandwidth management, to name a few. In mobile environments, data (educational curriculums in multimedia form) is usually distributed disparately across several sites. An interesting concept is modeling such an environment to find a way to allocate courseware (or educational curriculums in multimedia form) such that the probability for successful learning of the required courseware unit based on the learner’s learning behavior can be obtained. Assuming that learners need to access different curriculums from remote sites within the network environment, we can define the probability that the needed courseware unit be accessed and learned successfully based on all connected local and remote sites by determining all possible routing paths. In addition, arranging a suitable learning sequence for the learner is an important issue. We also consider time restrictions by helping learners meet the imposed time restrictions presents an additional challenge. In this thesis, we incorporate the concept of moving patterns and data allocation scheme to improve the probability of the required courseware unit that can be accessed and learned successfully over a mobile environment. A course planning scheme to suggest a learning sequence and to consider the time constrain problem is addressed. Specifically, the proposed approach uses two views in modeling: the physical view and the logical view. In the physical view, we consider and model the link reliability (communication links), Multi-access Probability (the access problems), moving patterns (user’s learning behavior), and data allocations (location management) in the mobile learning environment. In the logical view, we consider and model the course planning strategy to find the suitable way for a learner to learn a particular courseware unit. Numerical examples are given to enumerate the proposed approach. Several simulation results based different topologies are studied to illustrate the applicability of the proposed approach.