Here, I want to summarize my experiences and reading list that I use to find a problem and provide an abstraction about the way to solve it as a Ph.D. Thesis. In the first place I have started with some concepts as follow:
- Determininstic Networking
- NFV
- Optimization
- Reinforcement Learning
Then I choose to solve an optimization problem with reinforcement learning for considering the URLL funcions in NFV. At 23 Jan 2021 I had a meeting as an qualification exam for presenting my problem and its roadmap.
These points are discuss in the meeting:
- Problem definition must be S.M.A.R.T
- Consider more theoritical concepts in your research subject
- Access layer delay is important because you are mentioning the end-to-end delay
- Ultra Reliable and Low Latency Services concepts are useful
- Pakcet replication and reliability
- Cite https://ieeexplore.ieee.org/document/8458130
- Determininstic as an adjective is okay?
- Network slicing literature
The total grade of this meeting and the paper exam is 17.25.
These ideas are from myself and worth more investigation.
- Use veterbi algorithm on partially ordered chains
- Consider Shapers in Flow Control of the TSN and DetNet