The roles of associational and causal reasoning in problem solving

摘要

Efficiency and robustness are two desirable, but often conflicting, characteristics of problem solvers. This article describes the Generate, Test and Debug (GTD) paradigm, which combines associational and causal reasoning techniques to efficiently solve a wide range of problems. GTD uses associational reasoning to generate initial hypotheses, and uses causal reasoning to test hypotheses and to debug faulty hypotheses, if necessary. We contend that the characteristics of associational and causal reasoning differ mainly in the way they deal with interactions—associational reasoning presumes independence; causal reasoning explicitly represents interactions. This difference helps account for the strengths and weaknesses of the reasoning techniques, and indicates the problem-solving roles for which they are best suited. The GTD paradigm has been implemented and tested in several planning and interpretation domains, with an emphasis on geologic interpretation.