Collaborative Mechanisms Of Innate And Adaptive Computer Immune Systems

Computer Immune Systems, Which are also called Artificial Immune Systems, are biomimetic systems.These approachs simulates the principles of Biological Immune Systems, and they are adaptive defence systems. The Biological Immune Systems are made up of two layers, one is Innate Immune System and the other is Adaptive Immune System, the researches of Computer Immune Systems are also related to the two layers.Negative Selection Algorithm is the most popular model in Computer Immune Systems, the biological Theoretical basis of which comes from adaptive immune system. The essential of Negative Selection Model is to train lymphocytes with Self, delete those which can match Self and leave the ones which can detect Nonself. Negative Selection Model has the ability of detecting unknown Nonself, with features of adaptability and diversity, and these are consistent with the intelligent features of Biological Immune Systems. However, both the Self and Nonself sets are infinite and uncertainty, which cause the High computational complexity and high false recognition rate.Dendritic Cell Algorithm is a new computer immune algorithm, the biological Theoretical basis of which comes from innate immune system. The algorithm is not required to distinguish huge amounts of Selves and Nonselves, it is only necessary to concerned about whether the system is in danger or not. Although it can avoid the massive Self/Nonself processing problem, the key points such as the presentation of danger signals, the determination of danger states are still reliance on human experience, which makes the Dendritic Cell Algorithm less of adaption and diversity and losed the intelligence which a Computer Immune System should have.The Collaborative Algorithm of Innate and Adaptive Immune Systems in Computer Immune Systems is the latest computer immune algorithm, the biological Theoretical basis of which comes from the collaborative response mechanisms of Innate and Adaptive Immune System. The algorithm emphasizes the interaction between Innate and Adaptive Immune System and has many adavantages preserving features of adaptability and diversity, reducing the computational complexity and false recognition rate. However, due to the collaborative response mechanisms research of Innate and Adaptive Immune System in a relatively immature stage, there is now room for improvement in two tissues:how to express danger signals and how to establish collaborative mechanisms between Innate and Adaptive Computer Immune Systems.The thesis uses the latest immunology-danger theory to carry out the research of collaborative mechanisms between Innate and Adaptive Computer Immune Systems. Danger Theory considers that antigen causing the body at disk is the most important, and an antigen is self or nonself is not important. Antigen-presenting cells (Antigen Presenting Cell, APC) of the innate immune are responsible for the perception of danger signals and to capture antigens which are presented to the adaptive immune system, and lymphocytes of the adaptive immune system are responsible for the identification of antigens presented by APCs under the control of APCs, danger signal is a key factor of immune system activation.The main contents include the following components:Firstly, the thesis emphasizes the importance of Computer Immune Systems,then introduces the motivation of the thesis, including the limitations of existing Innate/Adaptive Computer Immune Algorithms, the advantages and limitations of the Innate and Adaptive Immune Systems collaborative algorithms, the point of the thesis.Secondly, the thesis describes the immune response knowledge,which is the immune theoretical basis of the thesis, including three types of immune response and sevaral types of immune response patterns, and introduces sevaral typical immue algorithms and their limitations.Thirdly, the thesis extracts and simplifies the dangerous theory of innate and adaptive coordination immune response mechanisms of danger theory, then proposes cooperative artificial immune model based on the innate and adaptive immune coordination response mechanisms of danger theory. The model consists of artificial APCs (danger perception layer, corresponding to the innate immune system), artificial Lymphocytes (immune response layer, corresponding to the adaptive immune system), and coordination mechanism between the two types of cells. APCs consisting of artificial danger perception layer are responsible for collecting and assessing information on abnormal changes which represent the danger level of the system and generate the guidance strategy of cell behavior to direct artificial Lymphocytes of the immune response layer to clear invading antigens; artificial Lymphocytes immune consisting of the immune response the invading antigens presented by artificial APCs according to the guidance strategy of cell behavior and generates the feedback adjustment strategy by assessing the response results to adjust the guidance strategy of cell behavior which is generated by aritifical APCs. There are two types of coordination mechanisms between two types of cells:including the guidance strategy of cell behavior which artificial APCs generate and the feedback adjustment strategy which artificial Lymphocytes generate.Then the thesis introduces the importance of how to express the signal in the artificial APC research, discusses the lack of signal expression method in the existing dendritic cell algorithm (DCA), and then introduces a new signal expression method. This method expresses different parameters change information of process level, host level, network level, and user-level into various danger signals through Differential and differential methods. The thesis improves the exsiting dendritic cell algorithm according to the need of the new signal expression method, and validates the advantage of the new signal expression method by comparing the performance changes when the new signal expression method is introduced into DCA before and after.Finally, the thesis builds an experimental system for collaborative model of Innate and Adaptive Computer Immune Systems. It firstly introduces the experimental system architecture and detailed implementation process, and then introduces the experimenta objectives to be accomplished, finally uses the rpc.statd data set to implement the testing experiments on the experimental system, including three experiments:parameters setting, validity and adaptability, and the experimental results are analyzed.This paper reaches the following conclusions:1) the proposed collaborative model of Innate and Adaptive Computer Immune Systems improves the behavioral strategies of artificial lymphocytes which can keep the consistency between artificial lymphocytes as the detector group and illegal antigens and improves the large-scale problem, false high rate and other problems, and the model has the good adaptability which is an important feature of Artificial Immune System;2)compared with the signal expression method in the existing Dendritic Cell Algorithm, the proposed new danger signal expression is a more effective expression for how to express abnormal change information of conputer system into danger signals.