智慧型系統在卵巢癌晶片之分子演化與控制---子計畫四---基於計算智慧之基因網路模型與穩定度之研究(II)

Abstract

本子計畫使用卵巢癌晶片資料，進而發展一些以計算智慧方法如混合GP-fuzzy neural 、混合 function Petri net、Bayesian神經網路與計算機圖論在基因網路模型之分析的建立，有關基因網路模型之建立我們採用二種方式，分別是非線性微分方程式數學模型與非微分方程式數學模型之建立，同時做基因網路之globally exponential 和 global asymptotic穩定度分析，藉由新方法的提出可改善及提供新工具在卵巢癌晶片之分子演化與控制之基因網路模型與穩定度之研究。另外以Matlab軟體開發個別基因網路數學模型之軟體與相關生勿物資訊視覺化軟體之整合開發

For this project, we propose several computational intelligence approaches such as hybrid genetic-programming-fuzzy-neural, hybrid function Petri nets, Bayesian neural networks, and graph theory to gene networks for ovarian cancer microarray data. Modeling methods of gene networks as a set of nonlinear differential equations, and as non-differential equations are proposed. Furthermore, the globally exponential and the global asymptotic stability analyses to the gene network will be carried out. That is, in this project, we will put forward and develop some novel and innovative approaches and tools for molecular evolution and control of ovarian cancer with microarray data using gene networks. Advanced software packages like Matlab will be used to develop computer programs for gene networks and visual software for bioinformatics.