A computational approach for linguistic systems and its application to hybrid control systems

Linguistic dynamic systems (LDS) are dynamic systems where words are used in place of numbers for computing and describing system behaviors. LDS are most suitable for modeling and evaluating high-level decision-making processes in which most of the required information and proposed policies are available only in a linguistic, not numerical, form. Evaluation results based on LDS models can provide fast and efficient analysis for complex systems, and can be used as coarse guideline to direct further search for more detailed and accurate solutions. This is extremely useful for intelligent hierarchical control systems, where high-level decisions are often made with imprecise linguistic terms and words for events and then refined into more specific and precise tasks in low-level units. The goal of this paper is to outline a computational theory for LDS that would enable us to conduct a global dynamic analysis, system design and synthesis for linguistic dynamic systems based on concepts and methods developed for conventional dynamic systems. An application of the proposed theory to hybrid control systems is discussed briefly.