Scheduling in the cloud computing infrastructure has a number of difficult problems, such as calculation time, budget, load balancing, etc. Several task scheduling techniques, such as GA and ACO, have been presented, and they have reportedly enhanced the performance of cloud datacentres with regard to a variety of scheduling criteria. The task scheduling problem is NP-hard because the number of solutions/combinations increases linearly with the issue's scale, such as the number of tasks and computer resources. Hence, fully and efficiently arranging user responsibilities is difficult. This paper proposes cloud computing metaheuristics and cluster-based load-balanced job scheduling. The suggested credits-based task scheduling algorithm, known as IGFCM-IFHO-EDQL, minimises the makespan, maximises resource utilisation, and adaptively minimises SLA violation by clustering all incoming jobs to the available Virtual Machines (VMs) in a load-balanced manner. Task-Length, Makespan, Task-Priority, Deadline, Degree of Inequality, and Cost are the six real-time criteria used to categorise cloudlets and virtual machines. The performance of the recommended task scheduling algorithm is evaluated in light of the most modern methodologies for scheduling tasks, and the results are shown here.

Cloud computing, cloud datacenters, task scheduling, load-balanced, credits, Impulsive Genetic Fuzzy c-Means and deep Q-learning algorithm