The foci of Cercia's research include adaptive optimisation and machine learning. Its core technical expertise include evolutionary computation, meta-heuristics and ensemble learning.
In adaptive optimisation, we are particularly interested in optimisation in a dynamic and uncertain environment, with or without constraints, and with single, multiple or many objectives. Using ensemble learning approaches, we are dealing with online learning, class imbalance learning and semi-supervised learning. In real world applications, we have worked on software engineering (e.g., software project scheduling, defect prediction, effort estimation, software module clustering, testing resource allocation, etc.), network optimisation (e.g., train re-scheduling, route optimisation, etc.), engine management systems (e.g., engine calibration using dynamic evolutionary optimisation), smart building (e.g., thermal modelling for buildings), circuit design (e.g., digital filter design), shape optimisation (e.g., turbine blades and car shapes), and materials modelling (e.g., aluminium alloy modelling).
The projects listed here represent only a subset of projects we are working on or have worked on. Feel free to make enquiries if you want to find out more.
- Theory of Evolutionary Computation: [Approximation | ECDONE | EADOP | DAASE | NICaiA]
- Ensemble Learning: [iSense | EPiCS | DAASE | SEBASE | MBC | NICaiA]
- Evolutionary Optimisation: [LSGO | Approximation | ECDONE | DAASE | Engine Control | EADOP | NICaiA]
- Search-based Software Engineering: [DAASE | SEBASE | NICaiA]
- Dynamic Optimisation: [ECDONE | EADOP | Engine Control | DAASE | NICaiA]
- Real-World Applications: [Engine Control | IEM]
- Co-evolution: [LSGO | ECDONE | NICaiA]