Discovering natural laws by Reduction

A polynomial algorithm, called Reduction, is presented to discover natural laws by analysing a set of experimental data, instead of a heuristic exploration which, when adopted in BACON, can only lead to rediscovering simple laws. A complex law with multiple variables involved can be discovered by reducing it to a binary law involving only two variables. Discovering a binary law is accomplished by a generate and test search. This search is so efficient that it does not need any backtracking and is able to cover most of possible laws. A reduction-based discovery system, called DISCOVER 2. 0 was developed with a flexible knowledge base and an ability of dealing with imperfect data. The system has been verified to be valid computationally practically, and theoretically, by discovering a great number of complex laws, and can be also viewed as a learning engine embodied in any intelligent systems to improve their performance by obtaining a general rule from the accumulated data.