基于主动推理的拟态云服务多余度裁决在线自适应优化算法OA

To address the problem that redundancy adjudication parameter configuration in mimic cloud services is diffi-cult to adapt to dynamic environments,and that security gain,service latency,and recovery effectiveness are difficult to optimize jointly,an active inference-based online adaptive optimization(AIOAO)algorithm was proposed.Firstly,for mimic cloud services,a quantitative mapping model between adjudication parameters and security gain,service latency,and recovery effectiveness was established,and the adjudication parameter configuration was formulated as a multi-objective online optimization problem in a dynamic and uncertain environment.Secondly,the AIOAO algorithm was pro-posed,in which an active inference mechanism was introduced to adaptively update adjudication parameters online.Ex-perimental results show that the AIOAO algorithm can dynamically adjust adjudication parameters according to system state changes,and improve security gain and recovery effectiveness while keeping service latency under control.