Currently, the cloud computing paradigm is experiencing rapid growth as there is a shift from other distributed computing methods and traditional IT infrastructure towards it. Consequently, optimised task scheduling techniques have become crucial in managing the expanding cloud computing environment. In cloud computing, numerous tasks need to be scheduled on a limited number of diverse virtual machines to minimise the imbalance between the local and global search space; and optimise system utilisation. Task scheduling is a challenging problem known as NP-complete, which means that there is no exact solution, and we can only achieve near-optimal results, particularly when using large-scale tasks in the context of cloud computing. This paper proposes an optimised strategy, Cuckoo-based Discrete Symbiotic Organisms Search (C-DSOS) that incorporated with Levy-Flight for optimal task scheduling in the cloud computing environment to minimise degree of imbalance. The strategy is based on the Standard Symbiotic Organism Search (SOS), which is a nature-inspired metaheuristic optimisation algorithm designed for numerical optimisation problems. SOS simulates the symbiotic relationships observed in ecosystems, such as mutualism, commensalism, and parasitism. To evaluate the proposed technique, the CloudSim toolkit simulator was used to conduct experiments. The results demonstrated that C-DSOS outperforms the Simulated Annealing Symbiotic Organism Search (SASOS) algorithm, which is a benchmarked algorithm commonly used in task scheduling problems. C-DSOS exhibits a favourable convergence rate, especially when using larger search spaces, making it suitable for task scheduling problems in the cloud. For the analysis, a t-test was employed, reveals that C-DSOS is statistically significant compared to the benchmarked SASOS algorithm, particularly for scenarios involving a large search space.

翻译：当前，随着分布式计算方法及传统IT基础设施向云计算模式的转变，该领域正经历着快速发展。因此，优化任务调度技术在管理不断扩张的云计算环境中变得至关重要。在云计算中，大量任务需要在数量有限且异构的虚拟机上调度，以最小化局部与全局搜索空间的不平衡，并优化系统利用率。任务调度是一个NP完全难题，这意味着不存在精确解，特别是在大规模云计算应用场景下，只能获得近似最优结果。本文提出一种融合莱维飞行的优化策略——基于布谷鸟的离散共生体搜索算法（C-DSOS），用于云环境中的最优任务调度以降低不均衡度。该策略基于标准共生体搜索算法（SOS），这是一种受自然启发的元启发式优化算法，专为数值优化问题设计。SOS模拟了生态系统中观察到的共生关系，如互利共生、偏利共生和寄生关系。为评估所提技术，采用CloudSim工具包仿真器进行实验。结果表明，C-DSOS优于任务调度问题中常用的基准算法——模拟退火共生体搜索算法（SASOS），尤其在处理更大搜索空间时展现出更优的收敛速率，使其适用于云计算任务调度问题。通过t检验分析显示，与基准SASOS算法相比，C-DSOS在统计上具有显著优势，特别是在涉及大搜索空间的场景中。