A Novel LC-LOA and S3R2GCNN-Based Dynamic Workflow Process Type Identification and Scheduling in Cloud

Workflow scheduling (WS) maps out the processes and manages the execution of interdependent works within a process. However, the existing studies did not identify the workflow process types for efficient WS. Therefore, this paper presents a novel LC-LOA and S3R2GCNN-based dynamic workflow process type identification and scheduling in the cloud. Primarily, the cloud users register and log in with the cloud server. Afterward, the user assigns the workflow, followed by attribute extraction. Subsequently, the hashcode is generated for attributes by using THA. Next, the workflow deduplication is checked. If the workflow is repeated, it is removed from the workflow pool; otherwise, the workflow is given for WPTIS. In WPTIS, the classifier is trained based on data acquisition, graph slicing, attribute extraction, feature extraction, feature selection by LC-LOA, process labeling by L-Fuzzy, and classification by S3R2GCNN. Also, the features are extracted from the cloud server and given to WS. Eventually, the workflow is scheduled by using the Linear Congruential Lyrebird Optimization Algorithm (LC-LOA). The results show that the proposed system obtained a high accuracy of 98.43%, outperforming conventional models.