本文分析了流媒体监控系统的特点,对视频压缩算法和监控网络进行了数学抽象。提出了流媒体监控系统中调度策略的非合作博弈模型,该模型能够反映各种流媒体监控系统的动态运行特征与相互竞争关系。根据调度方式与监控系统的特点设计了子系统的效用函数,并论证了该非合作博弈模型中Nash均衡点的存在性和唯一性。给出了调度优化问题的描述和基于遗传算法的解算方法。仿真和实验表明,该模型能够有效地反映子系统间的资源竞争,可以求得优化的带宽分配方案。
In this paper, characteristics of streaming media based surveillance system are analyzed, and an abstract model of video compression and network transportation is provided. Subsequently, the utility function of subsystems is designed, and the existence and uniqueness of Nash equilibrium in non-cooperative network scheduling game is proved. At last, an optimization method based on genetic algorithm is presented. Experiment and simulation results show that the non-cooperative model can effectively reflect the resource competition among subsystems, and the proposed method can be used to optimize the network bandwidth allocation.