利用基因演算法反算底質參數

Abstract

本文目的是以組合化的觀念，建立出一套反算海底底質地音參數的演算法，應用於匹配聲場處理，以估測海底底質的物理特性，將匹配聲場處理法視為組合最佳化問題，量測聲場(measurement field)勢必要與大量的計算聲場(replica filed)作匹配，本研究選擇基因演算法當作匹配聲場處理法的最佳化演算法。在算例中，以實際音傳損耗當作反算底質地音參數之目標，並進行匹配，以本研究建立之基因演算法進行底質地音反算，最後能夠推算出隨深度、距離改變之聲速、密度及吸收係數。

In the present study,we use the genetic algorithm,which is based on concept of Combinatorial Optimization Problem to invert the geo-acoustic parameters by matching acoustic filed data.While we consider the MFP as a optimization problem,it is unavoidable that a huge amount of replica field will be produced for matching with the measured filed data.So the simualated genetic algorithm was adopted as the optimization algorithm.In the study,transmission loss data and simulation are used to estimate bottom properties by Matched-Field-Peocessing scheme.The bottom properties are also estimated by genetic algorithm which include range-depent profile of sound speed,density and attenuation.