A Framework for an Evolutionary Computation Approach to Supporting Concept Generation

This paper proposes a framework for an evolutionary computation approach to supporting concept generation in engineering design. This approach attempts to assist designers in increasing quantity and diversity of design concepts. The framework integrates the Theory of Inventive Problem Solving (TRIZ) methodology into an evolution process. First, the product resources are analyzed. The relationships between the resources are constructed using Genetic Programming. Second, contradictions between the relationships are identified based on the physical features of the resources and their relationships. Finally, principles for resolving contradictions in the TRIZ matrix guide designers to generate alternative design solutions. A case study is conducted to show how the approach works.