情報工学的問題解決法を用いた局地降雨予測手法の開発に関する研究

Abstract

The prediction system for severe rainfall at the regional scale along a cold front is developed. We use GPV data and image processing method to detect cold fronts, which have meso a scale size. Wind distribution and variation of thermodynamic profile is able to be infered qualitatively with using the cold front detected and meteorological knowledge.We develop this inference system through the production system using the theory of Qualitative Reasoning, so called QR, which is one of the methods of AI-reasoning. Using these systems, severe rainfall that occurs in local area is predicted reasonably.