Resource management and knapsack formulation in distributed multimedia networks

In thus paper, the problem of allocating multiple finite resources to satisfy the quality of service (QoS) needs of multiple applications along multiple QoS dimensions is presented. A mathematical model that captures the dynamics of such adaptive problem is presented. This model formulates the problem as a multiple-choice multidimensional 0-1 knapsack problem (MMKP), an NP-hard optimization problem. A heuristic algorithm is then proposed to solve the MMKP. Experimental results demonstrate that our proposed algorithm finds 96% optimal solutions on average, and outperforms other heuristic algorithms for MMKP. Furthermore, the time required is on average 50% to 70% less than that required by other benchmark heuristics. These two properties make this heuristic a strong candidate for use in real-time multimedia applications.