Competitive randomized algorithms for nonuniform problems

Competitive analysis of dynamic power management strategies for systems with multiple power saving states

“…We compare the expected power dissipation of our online algorithm to that of the optimal offline algorithm to get a probabilistic competitive ratio. This method has been used in the context of two-state systems [5,6]. We generalize this work for multistate systems.…”

“…All of the previous work on competitive analysis for dynamic power management has concentrated on two-state systems [5,6,11]. We say that an algorithm is c-competitive if for any input, the cost of the online algorithm is bounded by c times the cost of the optimal offline algorithm for that input.…”

“…Earlier work on competitive analysis of dynamic power management strategies presented bounds on the quality of DPM solutions [11,5]. Competitive analysis has proven to be a powerful tool in providing a guarantee on the performance of an algorithm for any input.…”

“…Adaptive strategies, on the other hand, use the history of idle periods to guide the decisions of the algorithm for future idle periods. There have been a number of adaptive strategies proposed in the literature [4,5,1,2]. In particular, Chung, Benini and De Micheli [3] address multiple idle state systems.…”

A MAC layer power management scheme for efficient energy delay tradeoff in a WLAN

“…Chip-level power management features have been implemented in mainstream commercial microprocessors [11], [12]. Techniques for automatic synthesis of chip-level power management logic are surveyed in [13].…”