Adaptive Workload Scheduling in Multi-Cloud Environments using Machine Learning Techniques

Abstract

Multi-cloud environments provide improved scalability, reliability, and flexibility for modern
cloud-based applications. However, efficient workload scheduling across multiple cloud
platforms remains a challenging task due to heterogeneous resources and dynamic workload
demands. Traditional scheduling approaches often fail to adapt to changing system
conditions, resulting in inefficient resource utilization and increased task execution time. This
paper presents an adaptive workload scheduling framework for multi-cloud environments
using machine learning techniques. The proposed approach analyzes historical workload
patterns and system performance metrics to predict optimal scheduling decisions. By
intelligently distributing workloads across multiple cloud resources, the framework enhances
resource utilization and reduces task completion time. Experimental evaluation demonstrates
that the proposed machine learning-based scheduler improves workload balance and overall
system performance compared with conventional scheduling methods.