基于自适应遗传算法与集成学习的微服务异常检测方法  

Microservice anomaly detection method based on adaptive genetic algorithm and ensemble learning

作　　者：杨云皓 姜瑛 YANG Yunhao;JIANG Ying(Computer Technology Application Key Laboratory of Yunnan Province,Kunming 650500,China;Faculty of Information Engineering&Automation,Kunming University of Science and Technology,Kunming 650500,China)

机构地区：[1]云南省计算机技术应用重点实验室,云南昆明650500 [2]昆明理工大学信息工程与自动化学院,云南昆明650500

出　　处：《微电子学与计算机》2025年第3期49-58,共10页Microelectronics & Computer

基　　金：国家自然科学基金(62162038,61462049,60703116,61063006);国家重点研发计划(2018YFB1003904);云南省应用基础研究计划(2017FA033);云南省计算机技术应用重点实验室开放基金(2020101)。

摘　　要：当前微服务异常检测方法存在的问题主要包括仅使用单一数据源进行分析以及模型参数依赖于开发人员的经验。此外,一些方法不能有效应对样本不平衡问题,这些因素都会影响模型的准确率。为了解决上述问题,提出了一种基于自适应遗传算法与集成学习的微服务异常检测方法(Adaptive Genetic Algorithm and Ensemble Learning-Microservice Anomaly Detection,AGEL-MAD)。该方法首先通过特征组合的方式表示服务轨迹数据,并将其与性能数据进行融合。其次,为应对样本不平衡问题,引入了单分类支持向量机(One-Class Support Vector Machine,OCSVM)与多层感知机(Multi-Layer Perception,MLP)的集成学习。最后,为排除开发人员的影响,在异常检测过程中引入自适应遗传算法(AGA)以获取参数的最优组合,提高模型的适应性和准确率。通过在不同的数据集上进行实验,验证了AGEL-MAD方法的有效性。The main problems with current methods of microservices anomaly detection include the use of a single data source for analysis and the dependence of model parameters on the developer's experience.In addition,some methods can not effectively deal with the problem of sample imbalance,and these factors will affect the accuracy of the model.To address the aforementioned issues,a microservice anomaly detection method based on adaptive genetic algorithm and ensemble learning(AGEL-MAD)is proposed.Firstly,the service trace data is represented by feature composition,which is fused with the performance data.Secondly,in order to deal with the problem of sample imbalance,the ensemble learning of One-Class Support Vector Machine(OCSVM)and Multi-Layer Perception(MLP)is introduced.Finally,in order to exclude the influence of developers,adaptive genetic algorithm(AGA)is introduced in the anomaly detection process to obtain the optimal combination of parameters and improve the adaptability and accuracy of the model.The effectiveness of the method is verified by experiments on different datasets.

关 键 词：微服务 异常检测 样本不平衡 自适应遗传算法 集成学习 

分 类 号：TP306.3[自动化与计算机技术—计算机系统结构]