免疫计算的测不准有限计算模型

以人工免疫系统为研究对象,采用可信计算和系统论的方法,构建了免疫计算的测不准有限计算模型,将该模型划分为三层,即固有免疫计算层、适应免疫计算层和并行/分布计算层.这种免疫计算方法在数据库查询的基础上匹配了已知异体的特征,识别了已知异体的类型.此外,该方法在抗体规则匹配的基础上学习了未知异体,记忆了新异体的特征,在数据库中添加了新异体的特征.该模型通过毁灭器消除了异体,抑制了异体的复制和扩散,控制了免疫计算的毁灭机制的双刃性,监控了免疫计算的负载平衡,维护了系统的鲁棒性和安全性,克服了免疫学理论的不完备性的影响.实验结果表明,这种免疫计算机制检测了病毒和故障等异体,识别出该异体是"爱虫"病毒及其感染的文件,并消除了异体,抑制了异体的复制和扩散,确保了系统的安全性和鲁棒性.

Non-accurately-measurable limited computing model of immune computation

Artificial immune system was investigated using the approaches of trusted computing and systematics, and the non-accurately-measurable limited computing model of immune computation was built. Besides, the model had three tiers, i.e. the innate immune tier, the adaptive immune tier and the parallel/distributed computing tier. In the immune computation, the features of the anomaly were matched through the query in the database, and the type of the anomaly was recognized. Furthermore, unknown anomaly was learnt through the rule matching of the anomaly, the features of new anomaly were memorized, and the features of the new anomaly were added to the database. In the model, the anomaly was eliminated by the destroyer, the clone and proliferating of the anomaly were restrained, and the two blades of the destroying mechanism in the immune computation were controlled. Moreover, the load balance of the immune computation was monitored, the robustness and security of the system were maintained, and the imperfection of the immunology theories was overcome. The experimental results show that the immune computation mechanism detects the anomalies such as worms and faults, recognizes their types, i.e. the love worms and their infected files, and eliminates the anomalies, so that the clone and proliferating of the anomalies are restrained, and the security and robustness of the system are insured.