Developing Measures of Complexity for Engineering Design

Abstract

Developing objective measures for evaluating the complexity of the design problems would facilitate empirical studies that require use of “equivalent” but different design problems, development of design curriculums, and comparison of computer aided design automation tools. This paper discusses the similarities and differences between design problem complexity, design process complexity, and design artifact complexity. Three fundamental aspects to complexity are identified: the size, the degree of coupling, and the solvability. Alternative methods for measuring these characteristics of the design are discussed based upon computational views of complexity, information based views of complexity, and traditional design views of complexity. A method of measuring size as it relates to complexity is proposed for measuring the information content of design. A second method is proposed for decomposing a graph-based representation of design that provides a measure of the interconnectedness as it relates to complexity. Finally, two methods are proposed for determining the solvability complexity of design based upon the effort involved and the degree of freedom of design. These measures are developed specifically for parametric and geometric problems as found in embodiment design.