부산대학교 도서관

Multimedia content is distributed via all types of networks to viewers found in a variety of locations and using different types of devices. Increasing the performance of multimedia stream delivery requires overcoming many technological challenges, all of them having a direct effect on the user perceived quality of experience. The quality of experience influences in turn the quality of any e-learning process. This is as users in general and learners in particular are becoming increasingly quality-aware in their expectations. Therefore delivering a good quality video stream as a part of any e-learning process is very important. This paper proposes a new adaptive multimedia delivery algorithm which can be used in the context of e-learning. The Dynamic Quality Oriented Adaptation Scheme (D-QOAS) adapts the multimedia content sent based on both user preferences and network conditions, while adjusting dynamically its adaptation policy during delivery. Simulation results show that for different user profiles and various network conditions the improvement in end-user perceived quality is significant. Important benefits are obtained in terms of the total number of simultaneous users and in the link utilization, as well as in quality as measured by some video quality metrics.