Dynamic scheduling of independent tasks in cloud computing environment applying improved chicken swarm optimization and differential evolution

Cloud computing dynamically allocates computational resources to meet user demands, but task scheduling remains an NP-hard optimization problem due to multiple constraints such as task dependencies, execution time, and resource allocation. Metaheuristic algorithms have been widely used to tackle these challenges, yet existing approaches often suffer from premature convergence, poor exploration–exploitation balance, and inefficiencies in high-dimensional search spaces. To overcome these limitations, this study proposes an Improved Chicken Swarm Optimization with Differential Evolution (ICSODE), which integrates chaos-based adaptive parameter tuning to dynamically adjust the learning coefficient C for better convergence and Differential Evolution (DE)–based mutation and recombination to enhance population diversity and global search efficiency. A Dynamic Scheduling Framework (DSF.ICSODE) is also introduced, leveraging ICSODE to efficiently allocate independent tasks to virtual machines (VMs) in a cloud computing environment. The proposed ICSODE algorithm is first evaluated on CEC2017 benchmark functions, demonstrating superior mean error reduction compared to CSO, ICSO, ICSOCrossover, ICSOChaotic, and DE. In the second phase, ICSODE is applied within the DSF.ICSODE scheduling framework and tested on the NASA-iPSC parallel workload dataset, where it significantly reduces execution time by an average of 15.36% and improves response time by an average of 38.65%. Experimental results confirm that ICSODE outperforms existing CSO variants and DE in both function optimization and dynamic task scheduling, establishing it as a highly effective solution for cloud-based scheduling challenges.