AI-based End-to-End (E2E) Directional network slicing with guaranteed QoS over a highly dynamic network
The purpose of this research is to develop autonomously managed slicing of network resources in 5G/6G Networks that are expected to support many use cases ranging from autonomous vehicles, e-health, industry 4.0, entertainment, transport, smart cities etc. which will place a wide range of Quality of Service technical requirements 5G Network.
The 5G network is expected to consist of a mix of Macro, Micro and Pico cells within which communications (link and slice capacities), computational and storage resources will require to be located. Managing the deployment of mix of Macro, Micro and Pico cells and the communications, computational and storage resources to support these use cases will be require solving the set of all NP decision problems using a non-deterministic algorithm in polynomial time.
The objective of this research is to: 1. Study a range of use cases that will be required to co-exist and develop user, functional and technical requirements of 5G technical resources. 2. Develop the Problem model: A problem model is an abstract mathematical representation that captures the main characteristics of the problem to be optimised. Usually, models are intelligent simplifications of reality. It involves approximations/assumptions and sometimes may skip processes that are complex to represent mathematically but can easily be modified and is still able to provide useful insights to the modelled problem. 3. Develop the Problem formulation: Identify a set of decision variables, objective(s) and constraints that characterise the problem. 4. Develop the Optimisation Method: Once the optimisation problem is formulated, the next step is to autonomously solve the model using reinforcement learning, which finds the optimal values of the state, action and reward variable(s) to the model based on the objectives(s) and respecting the constraint(s) of the problem.
Typically, efficient algorithms are developed to solve the model, either to optimality or approximately. 5. Regardless of the meta-heuristic algorithm considered to solve a given optimisation problem, there are three core design questions common to all meta-heuristics in approaching an optimisation problem; the, definition of the states, actions and rewards that will guide the autonomous adaptation to the environment, and the definition of variation operators that move the algorithm from one point in the search space to another.
Studentships may be available depending on availability and research performance. Please enquire with Prof. John Cosmas for more details.
How to apply
If you are interested in applying for the above PhD topic please follow the steps below:
- Contact the supervisor by email or phone to discuss your interest and find out if you woold be suitable. Supervisor details can be found on this topic page. The supervisor will guide you in developing the topic-specific research proposal, which will form part of your application.
- Click on the 'Apply here' button on this page and you will be taken to the relevant PhD course page, where you can apply using an online application.
- Complete the online application indicating your selected supervisor and include the research proposal for the topic you have selected.
This is a self funded topic
Brunel offers a number of funding options to research students that help cover the cost of their tuition fees, contribute to living expenses or both. See more information here: https://www.brunel.ac.uk/research/Research-degrees/Research-degree-funding. The UK Government is also offering Doctoral Student Loans for eligible students, and there is some funding available through the Research Councils. Many of our international students benefit from funding provided by their governments or employers. Brunel alumni enjoy tuition fee discounts of 15%.