F E R M A T Formal Engineering Research using Methods, Abstractions and Transformations Technical Report No: 2003-13 Stochastic Modelling based approaches to deriving System Level Dynamic Power Management (DPM) strategies have been extensively studied in the literature. This paper extends our previous work on using probabilistic model checking to derive continuous Markov Chain based strategies for DPM, by applying similar techniques to discrete Time Markov Chain based strategies. This extension shows that probabilistic model checking provides a basis for a uniform framework for derivation, analysis and validation of stochastic Dynamic Power Management strategies with many additional advantages. This is a novel application of formal model checking of probabilistic systems in the area of system design. This approach allows us to obtain expected performance measures of the derived strategies by automated analytical means without expensive simulations. Moreover, one can formally prove various probabilistically quantified properties pertaining to buffer sizes, delays, energy usage etc., for each derived strategy. Comparison of various strategies under various stochastic behavioral assumptions can also be done formally without simulation. In this paper, we illustrate how we have implemented our uniform approach in the PRISM model checker framework and present results from realistic DPM scenarios based on a disk-drive example with multiple power management states. Formal Derivation and Analysis of Stochastic Dynamic Power Management Strategies based on Discrete-time Markov chain: A Probabilistic Model Checking Approach Gethin Norman, David Parker, Marta Kwiatkowska, Sandeep K. Shukla and Rajesh K. Gupta