一种异构实时系统中任务流均衡负载的算术模型

异构环境下如何提高系统自治并发处理的能力，是分布式系统研究领域的重要课题。而异构实时系统中系统自治并发处理的能力则体现在事件驱动任务流控制的资源最优分配。本文针对已有的资源分配策略的缺陷给出了一个通用的任务流均衡负载的算术模型，除了对任务流进行量化，使其可以动态分配以外，还可以在任务流发生变化时在满足系统最大利益的前提下最优化代价函数。另外，由于这个算术模型具有通用性，因此适用于各种资源分配的启发式搜索算法。

A Balanced Workload Mathematical Model in Heterogeneous Real-time Systems

How to improve the systems' concurrent processing capability in the heterogeneous environment has gotten the growing attention. The resource allocation of the event driving tasks flow in realtime systems can be used to de- pict the concurrent processing capability of realtime systems. Many resource allocation policies have emerged to get this purpose, however some defects still hamper the policies employing. One of them is how to quantify the workload and how to protect the systems' cost function without breaking the QOS constraints. We propose a balanced workload mathematical model to solve this problem. Besides this, BWMM's generality make it easily adapt to any heuristic re- search algorithm.