Semi-intuitive computations and artificial intelligence: Dimensional analysis as a case study

Abstract

The computational power of a semi‐intuitive or semiformal problem solving method is investigated using dimensional analysis as a case study. By modeling this problem solving method in a computer program, it is found that it provides a computational shortcut to the formal methods that are involved in solving the full equations of motion for a given problem by utilizing domain knowledge in the form of facts, typical behaviors, and anecdotal information about the physical system. This cooperation of knowledge of various forms seems to be a key characteristic of creative analysis and problem solving tasks.