Kuan, Yuan-HungYuan-HungKuanHu, Ming-CheMing-CheHuChen, Shien-TsungShien-TsungChen2025-12-012025-12-012024-01-0104921550https://www.scopus.com/record/display.uri?eid=2-s2.0-105007949633&origin=resultslisthttps://scholars.lib.ntu.edu.tw/handle/123456789/734259This study proposed a new prediction algorithm to forecast offshore significant wave heights collected from buoy stations during the typhoon warning period. The wave height and typhoon characteristic data served as the database to establish the input-to-output structure of the proposed forecasting model. The significant wave height is the main input parameter, and the remaining typhoon characteristics, such as the distance between the center of the typhoon and the buoy station, the central pressure of the typhoon, and the radius of the typhoon, are auxiliary parameters. The methodology, which includes the simple projection method, the convex combination, and a similarity index using the Euclidean distance, was developed to predict the value of the typhoon wave height in real time. Moreover, the simulated annealing method was used to find the optimal weight value for each model parameter to improve prediction results.falseSimplex projectionSimulated annealingWave height forecastingjournal article10.6937/TWC.202412_72(4).00042-s2.0-105007949633