Redesigning a Problem-Solver's Operators to Improve Solution Quality

摘要

The inability of a problem solver to produce solutions of a desired quality often may lie in the incorrect design of its operators. In this paper, we describe a method which, given a problem solver that produces a por solution for a given problem, and given the desired solution for it, uses a model of the problem-solver's processing and knowledge to identify faults in the specification of its operators and to appropriately redesign them. This method is based o nthe structure-behavior-function (SBF) model of problem solving that explicitly captures the functional semantics of the problem-solver's tasks, the compositional semantics of its problem-solving methods that combine the operators' inferences into the outputs of the overall task, its domain knowledge, and the "causal" interdependencies between its tasks, methods and domain knowledge. We illustrate and evalaute this learning method through AUTOGNOSTIC.