Co-evolution of the fitness function and design solution for design exploration

We use a Genetic Algorithm paradigm for modelling design exploration and distinguish this approach from design optimisation. The common assumption in design optimisation is that a fitness function is defined in advance. However, this is hardly the case for any practical design, especially for conceptual design. The fitness function changes and co-evolves with the generation of alternative design solutions. This paper presents an approach to the co-evolution of the fitness function and design solution by representing the fitness as part of the genotype. In this combined gene approach, the design solution part of the genotype is evaluated by the local fitness function part of the genotype. Through a two-phase crossover operation, both the fitness function and the design solution change with each generation. A example of designing a braced frame panel is used to illustrate this approach