In many areas of engineering, the development of optimal designs has become the search for a compromise between specifications that must be simultaneously satisfied. In general, the specifications that must be met are opposing, and so the conventional concept of optimal design becomes untenable. Problems become multi-objective optimisation problems when there is no single solution and there is a set of solutions in which each offers advantages over the others, but where none is better in every respect. The designer must finally choose one of the possible solutions according to the design preferences.
The CPOH group researches the development of new multi-objective optimisation algorithms based on artificial intelligence, new methods for the treatment of constraints, and new systems to support decision making.
The objectives of our research and development attempt to cover all the aspects that form a solution to a multi-objective problem:
- Development of new algorithms based primarily on evolutionary techniques, but without ruling out other approaches
- New systems of decision support for design engineering problems. Mechanisms that include designer preferences for methods of problem integration, as well as the development of tools to help select solutions
- Development of analysis and visualisation tools for multi-objective problem solving.