Main ongoing projects :

• Contributed book, Intelligent Control and Smart Energy Management : Renewable Resources and Transportation.
  

Click here for more info.

• Multi-objective optimization for multi-agent control strategies
  

The multi-objective concept applied to multi-agent systems affects the information exchange between agents, which invalidates main mathematical assumptions usually used to demonstrate algorithm convergence in multi-objective optimization for multi-agent systems, in which all objectives are prioritzed equally. Therefore, several concepts/theories are to be studied and demonstrated to ensure a reliable and safe deployment in practice. This includes proving the convergence to the optimal solution and the convergence limit for all time intervals. In addition, the evolution in real-time of the relative importance of objective functions and system topology changes are theoretical aspects that deserve exploration. Further theoretical development in this direction would open the way to new practical applications for which the importance of objectives are subject to change according to agents' tasks and the environment.

The theoretical development of this new class algorithms is promising for an abundant number of multi-agent system applications, e.g. unmanned aerial vehicles or teams of robots or may want to explore different regions of an area. The methods to be developed would be transferable from the field of robotics to energy, and allow prioritizing energy efficiency independently and intelligently. These methodologies are therefore likely to be applicable to ecological vehicles where their energy optimization is carried out according to several objectives such as the duration and conditions of the journey and the impact on the lifespan of the batteries up to intelligent control of smart grids to achieve better energy management.

• Soft computing development along with proof of convergence

The application of soft computing techniques is mostly based on empirical performance assessments and convergence is often not guaranteed, which are important drawbacks of such techniques. Therefore, this project includes devising algorithms based on soft computing techniques along with the study of convergence properties. The development of such techniques with theoretical proof will bridge the gap between empirical performance assessments and a rigorous mathematical understanding of soft computing algorithm properties, which includes theoretical and mathematical studies on soft computing properties such as proof of convergence, convergence rate, and computational complexity.

• Study of the common practices in optimization benchmarking

By studying common practices in optimization, this project aims to established guidelines and scientific criteria in optimization benchmarking.