engleski

Qualitative Modeling of Complex Systems Based on Similarity Calculus

The paper presents an efficient method for deriving qualitative correlative models of physical processes from observations of their behaviors. Most current techniques in qualitative modeling address learning a single qualitative differential equation (QDE) to model the system. Since many complex physical systems do not fall in this oversimplified class, we describe a novel technique that reads quantitative time-series data of process variables and promptly derives a qualitative correlative model. The technique is based on mapping quantitative time series data to binary coded qualitative difference vectors of process variables. The presented procedure significantly accelerates the modeling process. The results of applying the technique to a real-world thermodynamic system are shown.

Model design ; Process identification ; Qualitative modeling

nije evidentirano