Kubernetes Hybrid Scheduler This scheduler uses the NSGA-III algorithm to select the best nodes on which to place a batch of pods. Pods with three or fewer replicas, or those for which NSGA-III cannot find an optimal solution, are scheduled serially using the TOPSIS algorithm. Scheduler.py follows the Operator pattern to monitor the cluster state and system metrics. It relies on metrics-server, Prometheus Node Exporter, and iperf3 for data collection. Results: Increased availability, reliability, and scalability of the system Reduced node power consumption and lower communication delay Usage: Use KIND as a local Kubernetes cluster -> make buildkind Specify schedulerName: hybrid-scheduler and define custom resource requests in pod configurations (example).
