人工免疫系统:原理、模型、分析及展望

目前，受生物免疫系统启发而产生的人工免疫系统（Artificial ImmuneSystem,AIS)正在兴起，它作为计算智能研究的新领域，提供了一种强大的信息处理和问题求解范式，该文侧重以AIS的基本原理框架为线索，对其研究状况加以系统综述，首先从AIS的生物原型入手，归纳提炼出其仿生机理，主要包括免疫识别，免疫学习，免疫记忆，克隆选择，个体多样性，分布式和自适应等，进而对几种典型的AIS模型和算法分门别类地进行了细致讨论，随后介绍了AIS在若干具有代表性的领域中的应用情况，最后通过对AIS的特性和存在问题的分析，展望了今后的研究重点和发展趋势。

Artificial Immune System: Principle, Models, Analysis and Perspectives

Drawing inspiration from the vertebrate immune system, a new research field of Artificial Immune System (AIS) is springing up. As a novel branch of computational intelligence, AIS has strong capabilities of pattern recognition, learning and associative memory, hence it is natural to view AIS as a powerful information processing and problem-solving paradigm in both the scientific and engineering fields. This paper intends to give a comprehensive overview of AIS based on a preliminary theoretical framework, which is started with the brief interpretative introduction of biological models of vertebrate immune system, then followed with some extracted bionic principles, viz. immune recognition, immune learning, immune memory, clone selection, diversity generation and maintenance etc. The mapping from natural immune system to AIS models is emphasized in this paper. As a result, some typical AIS based models and algorithms are discussed through classifications. It is the real engineering applications that draw the broad attention of computer scientists to recognize the great potential of AIS, hereby some important application fields as information security, pattern recognition, optimization, machine learning, data mining, robotics, diagnostics and cybernetics etc. are reviewed. Then based on the property analysis of AIS, some key problems in the state-of-the-art of AIS research are investigated, through which we hope to gain deep insight into AIS and suggest some new ideas that may be of value for AIS model development. Finally, some possible research directions of AIS are given by the authors in a further step as the summary of this paper, among which the application of AIS to evolutionary design is emphasized.