一种多处理器全局实时调度中的任务负载最坏情况分配执行方式

Multiprocessor scheduling is known to be an NP-hard optimization problem, and no worst-case job arrival sequences have been identified for global scheduling real-time systems. In this paper, we present a novel idea to solve the problem of worst-case scenarios in global scheduling, and firstly introduce a hypothetical parallel executing system where any job except the studied one can be executed in parallel on all processors at any instant in time, we prove that such a hypothetical system leads to a worst-case scenario of any studied job's schedulability, therefore, the proposed results can be used for solving schedulability problems of multiprocessor global scheduling.