The development of cloud computing in current decades has led to it serving as the basis for a variety of systems. It enables customers to access a list of specified resources, act immediately and adaptably to customer preferences, and only be charged for actual utilization. One of the most important problems in cloud computing is Task Scheduling (TS). The issue is how to equitably distribute and organize the user-provided tasks for Virtual Machine (VM) execution. Also, user experience is directly impacted by the effectiveness of scheduling efficiency. As a result, the TS issue in cloud computing has to be more precisely addressed. In cloud computing, the TS is essential such that the optimal scheduling of task requests may boost network efficiency. The main objective of TS is to assign tasks to appropriate processors to create the shortest deadline achievable without compromising on priority criteria. Numerous research has been conducted to design TS schemes based on various metaheuristic and machine learning algorithms that satisfy several criteria such as minimization of the makespan, execution cost and energy. They have demonstrated that conventional TS is effective only to satisfy certain criteria and have devised an optimum solution using multi-objectives in cloud computing. This paper presents a systematic and extensive analysis of TS algorithms in cloud computing depending on the different optimization and machine learning algorithms. Also, it addresses the challenges in those algorithms and recommends a few possible solutions for improving the utilization of cloud computing.

Keywords : Cloud computing, Task scheduling, Virtual machine, Makespan, Metaheuristic, Machine learning, Optimization.