A design structure matrix approach for generating planning abstraction hierarchies

Research in Artificial Intelligence has been a forerunner in developing the most detailed and formalized theories that create consistent abstraction hierarchies for planning and problem solving. However, the representational methods to exploit these theories are complicated, which limit their application into many disciplines, specifically engineering. The objective of this paper is threefold: to simplify the representation of current AI-based planning, to identify the properties that ensure effective development of abstraction hierarchies, and accordingly, to develop a methodology for effective and consistent generation of abstraction hierarchies. The proposed methodology achieves these objectives by integrating the well-established AI hierarchical abstraction theories with Steward's practical Design Structure Matrices (DSM). The effectiveness of the developed methodology is demonstrated by applying it to the conceptual design on a facility layout of manufacturing plant that produces high voltage power cables.