Study on the energy-efficient resource allocation of distributed real-time systems

Scheduling algorithm based on dynamic voltage scaling may result in high extra energy consumption when taking frequency switching overheads into account.To resolve this problem,we present a heuristic resource allocation algorithm to even the tasks' utilization in distributed real-time systems using time Petri nets as specification models.In addition,by exploiting the rules of time Petri nets state space transformation,the underlying verification can be simplified,and then the efficiency of the resource allocation can be improved.Experimental results demonstrate that the proposed method can decrease frequency switching overheads and reduce energy consumption.