Advantages of Evolutionary Computation used for Exploration in the Creative Design Process

In early phases of design a wide exploration of the design space is crucial to the development of creative solutions. In this regard, Evolutionary Computation (EC), and in particular Genetic Algorithms, contain several qualities that can enhance exploration by opening the search process beyond the focus of finding a single "best" solution. Over the years many researchers in the area of creative thinking including Gordon, de Bono, Parnes, Osborn and others, have suggested design strategies that have interesting parallels in EC processes. For instance, a well known inhibitor of creative thinking is design fixation, where the suggestion of a particular solution makes it difficult to imagine other good solutions. Unlike many other computational search algorithms, EC methods work with populations of "fairly good" solutions. Therefore, there is less danger that creativity will be harmed by design fixation on one "best" solution. This paper shows through a specific example of a truss bridge how an EC based design exploration program can aid the designer by providing a selection of "pretty good" solutions rather than a single optimal solution. Other aspects of the EC program are also discussed including drawbacks to the method such as computational intensity as well as directions of future development.