Knowledge acquisition techniques for decision analysis using AXOTL and AQUINAS

Abstract

The effective application of current decision tree and influence diagram software requires a relatively high level of sophistication in the theory and practice of decision analysis.Research on intelligent decision systems aims to lower the cost and amount of training required to use these methods through the use of knowledge-based systems; however, application prototypes implemented to date have required time-consuming and tedious handcrafting of knowledge bases. This paper describes the development of DDUCKS, an "open architecture" problem-modeling environment that integrates components from Axotl, a knowledge-based decision analysis workbench, with those of Aquinas, a knowledge acquisition workbench based on personal construct theory. The knowledge base tools in Axotl can be configured with knowledge to provide guidance and help in formulating, evaluating, and refining decision models represented in influence diagrams. Knowledge acquisition tools in DDUCKS will allow the knowledge to be efficiently modeled, more easily maintained, and thoroughly tested. INTRODUCTION Progress in Automated Decision AnalysisThe Boeing Company has an urgent need for advanced automated decision support applications. Rapid growth in the complexity of strategic and tactical decisions has outstripped the capacity of conventional decision aids. (Langlotz, Shortliffe, and Fagan, 1986). Finally, knowledge-based system development environments do not generally provide facilities for integrating historical data with expert judgment (Spiegelhalter, Franklin, and Bull, 1990).One of the most promising approaches to dealing with decision complexity in a consistent, general-purpose manner is decision analysis (Howard, 1966;Keeney & Raiffa, 1976;Raiffa, 1968). In the past few years, researchers and developers have made important theoretical advances and have implemented several successful systems for automated support of the decision analysis process (reviews by Horvitz, Breese, & Henrion, 1988;Pearl, 1988;Neapolitan, 1990). Although a thorough discussion is beyond the scope of this paper, we wish to review three of the important developments that have led to the current state of the art. Following this, we will explain why we think that the development of automated knowledge acquisition tools is crucial to the future of efforts to deliver decision analysis methodology to a wider spectrum of decision makers and domains.Decision tree software. The development of decision tree software ( Figure 1) represented an important milestone in automated decision analysis support (Olmsted, 1982;McNamee & Celona, 1987). Through the use of general-purpose commercial tools, decision analysis techniques have become more widely known and used than ever before. At the same time,there are several drawbacks to the use of decision trees as a representation device. For one thing, they grow exponentially with problem size, making them impractical for problems of significant size. Additionally, the tree metaphor for decision problems often leads participants t...

From knowledge science to symbiosis science

“…1). Four years later, I was honored to join Brian, along with my mentor and friend John Boose (Boose, 1986;Boose and Bradshaw, 1987;Bradshaw and Boose, 1990;Bradshaw et al, 1991;Shema et al, 1990), in sending out a call for papers for a Workshop on Wisdom-Based Systems that was to take place on June 22-24, 1989 at the Rosario Resort on Orcas Island in Washington state. Here is a paragraph from that call:…”