An Queue learning-based scheduling strategy with Hybrid Lyrebird Falcon Optimization for load balancing-based cloud services

Cloud computing provides consumers access to computing resources whenever they need them without requiring them to be directly managed. Infrastructure as a service (IaaS) in this cloud model relies heavily on resource allocation and task scheduling. Although many methods and algorithms have been proposed to address the job allocation problem, effective scheduling remains a difficult study topic. Existing methods continue to have shortcomings despite continuous efforts. We propose a Hybrid Lyrebird Falcon Optimization (HLFO) algorithm in conjunction with a Q- Learning based Scheduling Strategy (QLSS) called QSHLFO to address this problem. An expanded version of the QSHLFO is also presented to address urgent situations. The proposed method reorders the processes in the queue according to the states of the virtual machines (VMs) affected by these processes. Deadlock situations, in which a virtual machine is unable to complete a task, are explicitly addressed by the enhanced QSHLFO, which ensures that the task is moved to another queue for execution. When the suggested approach is compared to current systems, the performance findings show that the QSHLFO schedules activities more effectively. Interestingly, QSHLFO outperforms current methods by distributing 500 jobs over 55 virtual computers in just 75s. The proposed method attained a makespan of 250, energy consumption of 8.5 J, resource utilization of 92.1 %, response time of 180ms, and throughput of 45 KBd.