Energy-efficient virtual machine placement in heterogeneous cloud data centers: a clustering-enhanced multi-objective, multi-reward reinforcement learning approach

Efficient virtual machine (VM) placement is vital for optimizing the performance of cloud data centers. While recent studies have addressed this challenge, many have overlooked the heterogeneity of cloud environments and the importance of scalability. This paper introduces a novel multi-objective algorithm designed specifically for VM placement in heterogeneous and large-scale cloud data centers. Our approach leverages the K-means algorithm to group VMs based on demand characteristics. Subsequently, a multi-reward reinforcement learning algorithm is employed to allocate these VMs to physical hosts. Despite its simplicity, the proposed method demonstrates exceptional efficiency. Simulation results reveal that our approach significantly outperforms established algorithms such as GMPR, GRVMP, FFD, NSGA-II, RLVMP, and BFD. Key performance metrics include the number of active devices, energy consumption, resource utilization (CPU and memory), VM migrations, and adherence to service level agreements, highlighting the superiority of our method.