甘蔗育種輔助系統之研製

Abstract

傳統的甘蔗育種程序須依賴經驗豐富的育種研究人員。目前，電腦輔助的 觀念運用很廣，例如，電腦輔助製造、電腦輔助設計、電腦輔助教學。農 業上的許多問題非常複雜，像甘蔗育種就是一個適合以電腦輔助的研究方 向。在這篇論文中，我們依據類神經網路架構及遺傳演算法，發展一套能 從台灣糖業研究所建立於1990的親本資料庫中，歸納出甘蔗交配模式的方 法。經由實驗，我們提出的方法其正確率可以達到70%，收集更多資料使 正確率適度提高且實際於交配圃實驗，是我們未來須加強的工作。 The traditional sugar-cane breeding process depends on the determination of experienced breeding researcher. It's well known that computer-aided methodology is very useful in many fields today, such as CAD, CAM, and CAI. Since the sugar-can breeding problem in agriculture field is complex, it seems that using computer-aided methodology is very suitable to solve this problem. In this thesis, we use the techniques of neural networks and genetic algorithms to construct a method in order to induce thesugar-cane cross model from the sugar-cane parent database which established by Taiwan Sugar Research Institute (TSRI) since 1990. The experimental results show that the correct percentagge for testing is about 70%. To collect more data to increase the correct percentage and experiment in the actual cross garden is the work that we must strengthen in the future.