Immunity and Information Sensitivity of Complex Product Design Process in Overlap Decomposition
University of New South Wales, Mechanical School, PhD Studen
University of New South Wales, Mechanical School, Senior Lecturer
Last modified: December 13, 2007
Structural decomposition of large scale systems into subsystems or agents has been a method of simplifying decision making for the complex systems. In most complex system designs the overlap decomposition of the systems is employed, which leads to overlap design tasks between design teams and more synchronous information transfer between them. A method for measuring the overall complexity of a system after being overlap decomposition is presented. It is shown that this measure is useful in deciding on an appropriate decomposition strategy which influences the design process efficiency. A combination of random or in some cases exhaustive search with spectral graph partitioning is proposed for extracting the decompositions with the dual objective of low overall complexity with maximized overlapping parts.