An adaptive biogeography-based optimization with cumulative covariance matrix for rule-based network intrusion detection

Biogeography-based optimization (BBO) is a well-known population-based metaheuristic optimization algorithm. The historical population distribution information is so underutilized that the global optimal value may not be searched efficiently in existing variants of BBO. To address the limitation, we propose in this paper a BBO framework based on cumulative covariance matrix (CCM) and dub it as CCM-BBO. The CCM operator is used in CCM-BBO to establish Eigen coordinate system and enables the migration operator independent on fixed coordinate system. By embedding CCM operator into BBO, the migration operator is implemented in the Eigen coordinate system to enhance the rotation invariance of BBO. To verify the performance of CCM operator, CCM-BBO has been applied to three state-of-the-art variants of BBO: SBBO, DE/BBO and EMBBO. The experimental results on 25 benchmark functions of CEC2005, 10 benchmark functions of CEC2020, and a real-world intrusion detection optimization problem have demonstrated that CCM-BBO is an effective framework to enhance the performance of BBO.

Duke Scholars

Author Feng Tian DKU Faculty