| Summary: | 碩士 === 國立雲林科技大學 === 電機工程系碩士班 === 94 === More and more multicast communications are becoming real-time. In real-time communications, messages must be transmitted to their destination nodes within a certain amount of time; otherwise the messages will be rendered futile. To support real-time multicast communications, computer networks have to guarantee an upper bound on the end-to-end delay from a source to each destination. This is known as the multicast end-to-end delay problem. On the other hand, if the same message fails to arrive at each destination at the same time, there will probably arise inconsistency or unfairness problem among users. This is related to the multicast delay variation problem. In the paper, we are concerned with minimizing multicast delay variation under multicast end-to-end delay constraint. The problem has been proven to be NP-complete. After its NP-completeness, many heuristic algorithms have been developed for it. However, these heuristic algorithms can just generate suboptimal solutions. In this paper, we will propose a mixed integer linear programming formulation to yield optimal solutions for this difficult problem. Computer simulations testify that compared with existing heuristic algorithms, our new mixed integer linear programming formulation can yield solutions with smaller multicast delay variation under different multicast end-to-end delay constraint.
In the paper we take Visual C++ as the development kit and compare CHAINS heuristic algorithm with DDVCA heuristic algorithm in some different items. CHAINS heuristic algorithm is to establish k-shortest paths from the source to each destination within tightest arrangement and recombination under different multicast end-to-end delay constraint. We know by the result of computer simulation, may obtain better delay variation compared to DDVCA heuristic algorithm. CHAINS heuristic algorithm is to discover an optimal path from the source to each destination in all paths, causes the delay variation to be smallest. DDVCA heuristic algorithm is to establish a multicast tree by CBT(Core Based Tree) to ask for the delay variation under different multicast end-to-end delay constraints. Although CHAINS heuristic algorithm has the better result, its routing path does not with an advantageous position of multicast. The routing way in all established paths from the source to each destination will affect the efficacy for unicast routing characteristics.
|
|---|
