This research aims at developing novel multi-objective evolutionary approaches using genetic algorithms and particle swarm optimisation.  The new approaches are tested on classical and real-world problems (e.g. Brain-Computer Interfaces and Cancer Chemotherapy).
Smart Multi-Objective Particle Swarm Optimisation using Decomposition (SMOPSOD) uses the decomposition approach proposed in MOEA/D and introduces a novel smart approach for sharing information among particles. SMOPSOD has been successfully customized and applied for solving channel selection for Brain computer interfaces with the collaboration of brain computer interfaces group in university of Essex. SMOPSOD is also applied for solving cancer chemotherapy problem, the optimisation aims at reducing the tumour size while maintaining a sufficient patient life quality.
Clustering based Framework for Leaders Selection in Multi-Objective Evolutionary Algorithms (CLS) is introduced as a novel framework for leaders selection in multi-objective evolutionary algorithms utilizing clustering techniques. Both objective and solution spaces are clustered so as each individual is assigned to two clusters; one in the objective space and the other in the solution space. An indirect mapping between clusters in both spaces is then defined to recognize regions with potentially better solutions.