The vital concern of this research work is to propose a new clustering algorithm based on IoT devices. This process is made up with four stages namely, Cluster Formation, Splitting and merging, CH selection and Data transformation. The first and foremost step is the cluster formation; in this k -mean clustering is exploited to cluster the network, where the count of needed clusters is optimally selected using a new algorithm. Further, this selection is to either split or merge the cluster structure. Subsequently, the proposed clustering model defines the optimal cluster head selection of each cluster by the proposed algorithm. Finally, in the data transmission phase, the delay optimized data fusion tree is subjected for reducing the manipulation of data fusion on transmission delay. More importantly, to solve all the optimization problems (optimal count of clusters and cluster heads), this paper introduces a new Cyclic Rider Optimization Algorithm (C-ROA), which is the modification of Rider Optimization Algorithm (ROA). To the end, the performance of the adopted method is compared over the other classical models like FF, GWO, WOA and conventional ROA in terms of delay, normalized energy, alive nodes and cost function measures.

中文翻译：

---

使用改进的骑手辅助集群的物联网最佳集群头选择

这项研究工作的重点是提出一种新的基于物联网设备的聚类算法。该过程由聚类形成，分裂和合并，CH选择和数据转换四个阶段组成。第一步也是最重要的一步。在这ķ 利用均值聚类对网络进行聚类，其中使用新算法最优选择所需聚类的数量。此外，此选择将​​拆分或合并群集结构。随后，所提出的聚类模型通过所提出的算法定义了每个簇的最优簇头选择。最后，在数据传输阶段，对延迟优化的数据融合树进行处理，以减少对传输延迟的数据融合操作。更重要的是，为了解决所有优化问题（集群和集群头的最佳数量），本文介绍了一种新的循环骑手优化算法（C-ROA），它是对骑手优化算法（ROA）的修改。最后，将所采用方法的性能与FF，GWO，