Search-Control Knowledge for the Interoperability Problem between Conceptual and Preliminary Structural Design

Abstract

In this paper, we discuss a computational framework for the conceptual and preliminary structural design process. The framework provides control facilities by reasoning tasks that work on a structural model. These reasoning tasks are integrated into the design process domain by encapsulation into primitive tasks from an hierarchical task network. The initial task network is executed in the same order as planned for and the current design state is taken into account for controlling the compositional design task. Thus, the overall complexity of the design process can be reduced to cope with realistic structural design problems. First, we categorize the conceptual and preliminary structural design process in terms of its tractability by methods from Artificial Intelligence. Thereby, we define the various knowledge types and methods that are used by practitioners as common sense knowledge. We formulate on this basis a problem for conceptual and preliminary structural design, which respects the former introduced knowledge types that are applicable to solve the problem. Second, we specify the constituents of the conceptual and preliminary structural design process on the background of a F-B-S design model and assign inference procedures that accomplish the envisaged functionality in the structural design domain. Third, we introduce the knowledge level for control in order to analyse carried out design studies in regard to available methods for the representation of the found knowledge types. They enable reasoning tasks on the structural evolutionary model. Fourth, the reasoning tasks are defined in a logical framework. The tasks are finally presented along a realistic design sequence and feature their contribution to the overall control of the design process comprised of different sub-problem design spaces.