In: Proc. 37th Design Automation Conference (DA'2000), 5-9 June 2000, Los Angeles, CA, pages 352-356. 2000.
Abstract: This paper introduces a new technique for modeling and solving the dynamic power management (DPM) problem for systems with complex behavioral characteristics such as concurrency, synchronization, mutual exclusion and conflict. A power-managed distributed computer system is modeled as a controllable generalized stochastic Petri net (GSPN) with cost. The obtained GSPN model is automatically converted to an equivalent continuous-time Markov decision process. Given the delay constraints, the optimal power management policy for requests for system components as well as the optimal dispatch policy for requests are calculated by solving a linear programming problem based on the Markov decision process. Experimental results show that the proposed technique can achieve more than 20 percent power saving compared to other existing DPM techniques.
Keywords: Markov decision processes, concurrent systems, dynamic power management, stochastic Petri nets.