Rectangular phase-portrait approximation based on qualitative reasoning
基于定性推理的矩形phase-portrait近似

The core of the rectangular phase-portrait approximation is the efficient partition of the control model. The phase-portrait approximation based on quality reasoning is proposed. An approach for mode partition is then presented based on the characteristic of the vector field, interesting polynomials and their Lie-derivative. A method for the refinement of the abstract model based on the refined polynomials is also given. Experiment shows that the phase-portrait approximation based on the qualitative-reasoning partition obviously reduces the partition number of the mode state space, and enhances the verification efficiency.