A Cost-Driven Intelligence Scheduling Approach for Deadline-Constrained IoT Workflow Applications in Cloud Computing

Cloud computing is a potent platform for delivering high-quality computational services to intricate IoT applications. However, effective scheduling approaches are essential to meet application demands while maximizing cloud computing’s potential. In this study, we propose an innovative workflow scheduling method for addressing the cost-effective, deadline-constrained scheduling challenge of IoT applications in cloud computing systems. Our solution, the F-ACO algorithm, leverages a hybrid intelligence approach that combines Ant Colony Optimization (ACO) with a cost-driven heuristic strategy. The primary goal is to minimize workflow scheduling costs while ensuring that workflow deadlines are met. F-ACO introduces a deadline distribution method to derive task sub-deadlines, enabling dynamic adjustments for unscheduled tasks to meet workflow deadlines. Furthermore, we introduce an adaptive ACO-based task ordering mechanism with self-adaptive heuristic information to optimize task scheduling sequences, reducing search space redundancy and enhancing convergence speed. The approach includes a cost-driven task scheduling method designed to allocate each task to a virtual machine with minimal execution cost and idle time, further optimizing the overall workflow scheduling cost. To validate our F-ACO algorithm, we conducted numerous simulations using real-world workflows and compared its performance against state-of-the-art algorithms. Our experimental results affirm F-ACO’s competitive edge in effectively scheduling IoT applications in cloud computing environments.