The state of the art in designing engineering structures depends to a large degree on procedures and assumptions that date back to the 19th century. That time saw the advent of a new construction material with considerable tensile strength, namely steel. In the wake of and in combination with this innovation new statical calculation methods gave rise to buildings of hitherto unknown height or free span. These calculation procedures rest on a categorization of statical systems that still influences present day thinking. More or less all effective load bearing structures exhibit a high degree of regularity. They usually have one or more axes of symmetry. Some utilize arch or catenary shapes, but most consist of a large number of similar or identical elements. The structure’s geometry was and is heavily influenced by what engineers are able to handle and thus to calculate efficiently.
The recents developments in the field of civil engineering will help to overcome the above-mentioned limits:
First: advances in soft- and hardware make it possible to easily and swiftly assess a structure’s statical properties.
Second: State-of-the-art optimization procedures that are partly inspired by nature allow one to search large solution spaces in reasonable amounts of time.
Third: computer-aided fabrication methods facilitate the economic production of customized structural components.
These three developments can be combined to generate structures with intrinsic load-bearing mechanisms. Their design contradicts the traditional conception of an engineering approach which consists of an a priori selection of the main structural components according to external boundary conditions. Instead the statical system emerges from the design process according to intrinsic principles of the optimization algorithm and the given boundary conditions.
There is a high density of individuals in the vicinity of the best solution so far. Given enough computational resources the globally best solution will eventually emerge from the optimization process.