Genetic algorithm-based multi-objective optimisation and conceptual engineering design

In this paper we present a genetic algorithm based system for conceptual engineering design. First, we present a method based on preference relations for transforming non-crisp (qualitative) relationships between objectives in multi-objective optimisation into quantitative attributes (numbers). This is integrated with two multiobjective genetic algorithms: weighted sums GA and a method for combining the Pareto method with weights. Examples of preference relations application together with traditional genetic algorithms are also presented. A further method for dynamical inclusion and modification of extra constraints (not included in the mathematical model of the system) via scenarios is presented. Its use is discussed and potential applications indicated. Finally, some future work paths are mentioned.