The Adaptation of Structural System Designs Using Genetic Algorithms

The case-based reasoning paradigm has been widely used to provide computer support for recalling and adapting known designs to novel design situations. Case-based reasoning adaptation algorithms generally rely on domain knowledge and heuristics in order to change past designs to solve new problems. Unfortunately, the amount of domain-specific knowledge and domainand task-specific heuristics needed for such case adaptation is usually quite large. In contrast, in our current work we are exploring the use of a generalpurpose, knowledge-lean method based on genetic algorithms for the subtask of case adaptation. We are developing the methodology for the conceptual structural design of tall buildings.