Combining dimensional analysis and heuristics for causal ordering.

Shen, Qiang; Peng, Taoxin

Abstract

This paper is a refined version of the work that the authors presented at the 13th International Workshop on Qualitative Reasoning, jointly with the late Dr Rob Milne. It is dedicated to Rob in recognition of his significant contribution and support for the research described herein.

The paper presents a novel approach for generating causal dependencies between system variables, from an acausal description of the system behaviour, and for identifying the end causal impact, in terms of whether a change in the value of an influencing variable will lead to an increase or a decrease in the value of the influenced variables. This work is based on the use of the conventional method for dimensional analysis developed in classical physics, in conjunction with the exploitation of general heuristics. The utility of the work is demonstrated with its application to providing causal explanation for a benchmark problem that involves a dynamic feedback loop. The results reflect well the common-sense understanding of the causality in such a system that is otherwise difficult to capture using conventional causal ordering methods.

Citation

Shen, Q., & Peng, T. (2006). Combining dimensional analysis and heuristics for causal ordering. In A. Bundy, & S. Wilson (Eds.), Rob Milne: A Tribute to a Pioneering AI Scientist, Entrepreneur and Mountaineer (39-53). IOS Press

• Publication Date: 2006-08
• Deposit Date: Mar 16, 2010
• Peer Reviewed: Peer Reviewed
• Volume: 139
• Pages: 39-53
• Book Title: Rob Milne: A Tribute to a Pioneering AI Scientist, Entrepreneur and Mountaineer
• ISBN: 1-58603-639-4
• Keywords: Rob Milne; causal dependencies; system variables; influence; dimensional analysis; heuristics
• Public URL: http://researchrepository.napier.ac.uk/id/eprint/3428