Formal notations for system performance modeling need to be equipped with suitable notations for specifying performance measures. These companion notations have been traditionally based on reward structures and, more recently, on temporal logics. In this paper we propose an approach that combines logics and rewards, together with a definition mechanism that allows performance measures to be specified in a component-oriented way, thus facilitating the task for non-experts. The resulting Measure Specification Language (MSL) is interpreted both on action-labeled continuous-time Markov chains and on stochastic process algebras. The latter interpretation provides a compositional framework for performance-sensitive model manipulations and emphasizes the increased expressiveness with respect to traditional reward structures for implicit-state modeling notations.
Mixing Logics and Rewards for the Component-Oriented Specification of Performance Measures
Aldini, Alessandro;Bernardo, Marco
2007
Abstract
Formal notations for system performance modeling need to be equipped with suitable notations for specifying performance measures. These companion notations have been traditionally based on reward structures and, more recently, on temporal logics. In this paper we propose an approach that combines logics and rewards, together with a definition mechanism that allows performance measures to be specified in a component-oriented way, thus facilitating the task for non-experts. The resulting Measure Specification Language (MSL) is interpreted both on action-labeled continuous-time Markov chains and on stochastic process algebras. The latter interpretation provides a compositional framework for performance-sensitive model manipulations and emphasizes the increased expressiveness with respect to traditional reward structures for implicit-state modeling notations.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
