697 research outputs found
Heuristics for ioco-Based Test-Based Modelling (extended abstract)
No full textItem does not contain fulltextFMICS 200
Computing structural properties of symmetric nets
No full textStructural properties of Petri Nets (PN) have an important role in the process of model validation and analysis. When considering Stochastic PNs, comprising stochastic timed and immediate transitions, structural analysis becomes a fundamental step in net-level definition of probabilistic parameters. High Level PN (HLPN) structural analysis still poses many problems and is often based on the unfolding of the HLPN model: this approach prevents the exploitation of model behavioural symmetries. A more effective alternative approach consists in providing a language, along with an associated calculus, making it possible to derive expressions defining structural relations among node instances of a HLPN model in a symbolic and parametric form: this has been proposed in the literature for Symmetric Nets (SN). The goal of the present paper is to summarize the language defined to express SNs’ structural relations and to formalize the derivation of a basic set of such relations; in particular the algorithms to compute the Structural Mutual Exclusion relation and the symmetric and transitive closure of Structural Conflict are an original contribution of this paper. Examples of applications are also included. The algorithms required to support the calculus for symbolic structural relations computation have been recently completed and implemented in a tool called SNexpression
Implementing Symbolic Models for Value Passing in TwoTowers
No full textWe describe an extension of TwoTowers, a software tool for the functional and performance analysis of concurrent and distributed systems modeled in EMPA, in which a symbolic model based support for data driven computations is implemented and we recall its advantages both from the expressiveness and the analysis standpoint
Uniting Academic Achievements on Performance Analysis with Industrial Needs
No full textItem does not contain fulltex
Parallel CSRL Model Checking:First Results and Pointers to the Future
No full textThe verification of quantitative aspects of a system, like performance and dependability, by means of model checking has become an important and vivid area of research over the past decade. Recently, great progress has been made in the combined analysis of a system's performance and dependability (the so-called performability) by means of model checking. Both the logic CSRL (continuous stochastic reward logic) and a number of model checking algorithms for this logic are tangible outcomes of this research. To evaluate CSRL properties, however, it is necessary to solve large systems of partial differential equations (PDEs). The inherent time complexity of the model checking algorithms makes CSRL model checking for system models with more than 100,000 states practically infeasible. To overcome these difficulties, we have investigated various possibilities to parallelise the CSRL model checking algorithms. First practical experiments with a parallel version of the so-called path-exploration-based algorithm will be reported
MANOVA modelling of a chiropractic longitudinal study using multiple imputation
No full textThe purpose of this report is to present the detailed statistical analysis of a randomised, placebo-controlled trial comparing two different treatment modalities to an intervention of no known benefit for people with acute or subacute thoracic spine pain.
The therapy arms consist of Spinal Manipulative Therapy (SMT) and Graston Technique (GT) and the placebo is a non-functional ultrasound. A placebo group was utilised because at present there are no proven treatments for non-specific thoracic pain. This trial is registered with the Australia and New Zealand Clinical Trials Registry. Ethics approval has been granted by Murdoch University Human Research and Ethics Committee, number 2007/274.
The aim of this three arm trial was to test the efficacy of SMT and GT as independent modalities compared to detuned ultrasound for the outcomes of pain and disability. The latter were measured using the Visual Analogue Scale (VAS) and a modified Oswestry Back Pain Disability Index. The study was conducted at the Murdoch University Chiropractic student clinic in Perth, Australia, and the protocol published in Crothers et al (2008).
In this report, Section 2 provides an initial exploratory analysis of the data, Section 3 outlines the statistical models used in the final analysis, Section 4 defines these models in mathematical terms, Section 5 discusses the management of missing values via multiple imputation and Section 6 presents the results of the statistical modelling and hypothesis tests. The clinical study will be published in full elsewhere
Performance measures other than time
No full textThis presentation shows a few possible performance measures that might be interesting and possible evaluation methods
- …
