Abstract: Task-Resource mapping is the biggest challenge in cloud environments, especially when dealing with dynamic workloads and multi-objective constraints such as lowering cost, time, and maximizing utilization. A solution based on a metaheuristic-based multi-objective cloud scheduling (MMHEMOCS) is proposed in this paper. First, a novel integration of the crow search algorithm with electric fish optimization is proposed with dynamic parameters. This proposed dynamic CSA-EFO is further designed for multi-objective scheduling, and a new fitness function is also designed. This proposed framework is simulated across homogeneous and heterogeneous cloud environments. The experimentation is conducted on a set of resources with configuration settings as per the environments, with a set of different tasks. The performance metrics, such as energy consumption, resource utilization, cost, and execution time, are used to assess the proposed method. The comparisons with different optimization algorithms and their hybrid approaches are also performed. All the comparison results have shown the efficacy of MMHEMOCS for both homogeneous and heterogeneous environments. Also, the comparison of the proposed framework for homogeneous and heterogeneous environments has shown that the proposed algorithm handles both environments very well and performs better.

Key words: scheduling, cloud, homogeneous, heterogeneous, dynamic, optimization, multi-objective