Applying successive improvement transformations for intelligent problem solving

Abstract

An improvement transformation is a transformation which has the following two properties, namely, (1) after applying it, all the already achieved subgoals are still being accomplished and (2) additional new subgoals or partial new subgoals may be accomplished. By using improvement transformations, an intelligent problem‐solving method is presented with illustrative examples. This method consists of three parts: (1) goal decomposition and subgoal ordering; (2) successive improvement transformations; and (3) backtrack algorithms. The intelligent problem‐solving method using successive improvement transformations may have useful applications in goal decomposition, goal regression, genetic algorithms, robotics, and future expert system design.