A method for acquisition of QoS mapping rules for distributed multimedia applications

The ability to guarantee real-time delivery and satisfactory picture quality is mandatory when transmitting video over distributed multimedia applications. Quality-of-service (Qos) management has recently emerged as a critical research area. Network and terminal resources must be ensured in advance in order to deliver video that guarantees Qos, and the amount of these resources are calculated based on the Qos that is requested for the video. Different concepts of Qos apply to different system levels, thus requiring conversion of Qos parameters (called Qos mapping) between the higher application level and lower network and terminal levels. In the past, system designers have written rules in advance to effect Qos mapping, but so far, no effective descriptive method has been proposed. This paper addresses this lack by proposing a method of autonomously acquiring Qos mapping rules based on reinforcement learning, and computer simulation is performed to demonstrate that the acquired Qos mapping rules can satisfy the quality reinforcements of the user.