膜计算的非监督学习模型与机理研究

项目名称： 膜计算的非监督学习模型与机理研究

项目编号： No.61472328

项目类型： 面上项目

立项/批准年度： 2015

项目学科： 计算机科学学科

项目作者： 王军

作者单位： 西华大学

项目金额： 82万元

中文摘要： 从细胞生物学的各种机制和特征所启发的众多膜计算模型（也称膜系统）为计算机科学提供了一种新颖的并行计算理论、方法和技术。当前，严重地制约膜计算理论在实际工程应用的关键问题之一是这些膜计算模型普遍缺乏学习能力。膜计算的学习问题是膜计算领域亟待解决的一个公开困难问题。针对膜计算的非监督学习问题，拟从膜计算学习模型的构建出发，着力解决膜计算模型的学习机理这一关键问题，在此基础上，依据不同聚类问题的性质和特征，发展出若干高效的分布式并行聚类算法，以及对学习模型和算法进行测试、验证与评价，实现项目的研究目标。这个科学问题的解决将为膜计算应用提供一类新颖的模型和算法，同时也为计算机科学高性能计算的发展提供新的思路和途径。

中文关键词： 膜计算；非监督学习；聚类；模型；机理

英文摘要： Membrane computing models (known as membrane systems), inspired from various mechanisms and characteristics of cell biology, provide a novel parallel computing theory, method and technology for computer science. Now, one of the key questions which have severely restricted application of membrane computing theory to practical engineering is that these membrane computing models generally lack learning ability. The learning problem of membrane computing is an open-known difficult which needsto be solved urgently. Aiming at unsupervised learning problem of membrane computing, we will deeply research and explore their learning mechanism from starting to construct the learning models of membrane computing. On this basis, several efficient distributed parallel clustering algorithms are developed according to the properties and characteristics of different clustering problems. Furthermore, the learning models and algorithms are tested, validated and evaluated in order to achieve the research objectives. The solutions to the scientific problem in this project will not only provide a class of novel learning models and algorithms for membrane computing, but also give new ideas for and approaches to the development of high-performance computing of computer science

英文关键词： Membrane Computing;Unsupervised learning;Clustering;Model;Mechanism