Machine-learning-based self-adaptation of component ensembles
Thesis title in Czech: | Adaptace komponent a jejich kooperace na základě strojového učení |
---|---|
Thesis title in English: | Machine-learning-based self-adaptation of component ensembles |
Key words: | adaptace|strojové učení|komponenty|kooperace |
English key words: | self-adaptive|machine learning|components|ensembles |
Academic year of topic announcement: | 2021/2022 |
Thesis type: | diploma thesis |
Thesis language: | angličtina |
Department: | Department of Distributed and Dependable Systems (32-KDSS) |
Supervisor: | prof. RNDr. Tomáš Bureš, Ph.D. |
Author: | Mgr. Michal Töpfer - assigned and confirmed by the Study Dept. |
Date of registration: | 09.11.2021 |
Date of assignment: | 09.11.2021 |
Confirmed by Study dept. on: | 04.01.2022 |
Date and time of defence: | 15.06.2022 09:00 |
Date of electronic submission: | 03.05.2022 |
Date of submission of printed version: | 16.05.2022 |
Date of proceeded defence: | 15.06.2022 |
Opponents: | doc. RNDr. Pavel Parízek, Ph.D. |
Guidelines |
Modern smart systems, such as IoT (Internet of Things) and CPS (Cyber-Physical Systems) are typically required to dynamically self-adapt and reconfigure their structure based on the situation in the environment. To manage the complexity and describe the dynamically evolving architecture of such systems, architectural models featuring cooperating components have been introduced. One important representative of these approaches are the autonomic component ensembles [1], which ensure the coordination among the components which pursue a same goal. The ensembles are formed and dismantled dynamically based on the changes of the properties of the components and they can overlap and be nested. The formation of the ensembles was originally formulated as a constraint optimization problem. Later, a machine-learning-based solution was proposed to improve the ensemble resolution [2].
The goal of this thesis is to further integrate the machine learning algorithms with the ensemble resolution process. In particular, the thesis will focus on how machine learning can be used to predict values that the self-adaptive systems base their adaptation decisions on. This will allow proactive self-adaptation based on architectural description with ensembles. The predictions should be transparent for the ensemble resolution process so the predicted properties can be used for the resolution as if they were normal properties of the component. Further, the approach shall be aligned with the architectural perspective of autonomic component ensembles. |
References |
[1] Bures, T., Gerostathopoulos, I., Hnetynka, P. et al. A language and framework for dynamic component ensembles in smart systems. Int J Softw Tools Technol Transfer 22, 497–509 (2020). https://doi.org/10.1007/s10009-020-00558-z
[2] Bureš T., Gerostathopoulos I., Hnětynka P., Pacovský J. (2020) Forming Ensembles at Runtime: A Machine Learning Approach. In: Margaria T., Steffen B. (eds) Leveraging Applications of Formal Methods, Verification and Validation: Engineering Principles. ISoLA 2020. Lecture Notes in Computer Science, vol 12477. Springer, Cham. https://doi.org/10.1007/978-3-030-61470-6_26 |