Titulo Estágio
Building Run-time Probabilistic Verification into Architecture-Based Adaptive Systems
Área Tecnológica
Engenharia de Software
Local do Estágio
DEI
Enquadramento
The provision of assurances for self-adaptive systems presents its challenges, since uncertainties associated with their operating environments often hamper the provision of absolute guarantees that system properties can be satisfied. There are existing development-time approaches for the evaluation of self-adaptive systems to provide levels of confidence regarding trustworthy service delivery in spite of changes in the system and/or its environment. However, development-time evaluation has limitations due to the dynamic nature of self-adaptive systems, whose behavior depends on run-time conditions that continuously change.
Objetivo
In a nutshell, the objective of this internship is endowing an architecture-based self-adaptation platform (in this case, Rainbow, developed at CMU) with probabilistic verification capabilities at run-time that enable on-the-fly quantification of levels of confidence on the satisfaction of specified system properties, reasoning in terms of probabilities.
Plano de Trabalhos - Semestre 1
1. Analyzing the architecture and implementation of the Rainbow Framework for architecture-based self-adaptation (1 month).
2. Detailed description and discussion of the challenges posed by the incorporation of run-time probabilistic verification at run-time (3 weeks).
3. Analysis and discussion of candidate probabilistic-model checking tools to be integrated in Rainbow (e.g., PRISM, MRMC, etc.) (3 weeks).
4. Preliminary design of a new architecture for the Framework including run-time probabilistic verification. (1 month).
Plano de Trabalhos - Semestre 2
1. Implementation of a prototype of the new version of the framework. (4 months).
2. Validation of the prototype (1 month).
3. Writing final report (2 months).
*The activities described in the different semesters may overlap in time.
Condições
Work to be developed at SSE/CISUC
Orientador
Marco Vieira
mvieira@dei.uc.pt 📩