Performance Trees: Expressiveness and Quantitative Semantics

Abstract

Performance trees are a recently-proposed mechanism for the specification of performance properties and measures. They represent an attractive alternative to stochastic logics, since they support traditional stochastic model checking queries, while also allowing for the direct extraction of a wide range of quantitative measures. In this paper we illustrate differences in expressiveness between performance trees and continuous stochastic logic (CSL), and present quantitative semantics showing the mathematical basis underlying performance tree operators. As a running example, we demonstrate performance query specification with performance trees on a stochastic Petri net model of a healthcare system.