Selecting Key Feature Sequence of Resource Services in Industrial Internet of Things

Abstract

The Industrial Internet of Things (IIoT) achieves reliable operation of physical resources through real-time perceiving. In a smart factory, it provides a new way to share manufacturing resources as services or to use resource services of other enterprises. However, selecting a sequence of resource services, called the resource service chain (RSC), for a collaborative business process is difficult because the correlated resource services are usually selected independently by different organizations. The features of resource services, as an important intrinsic factor, should be used to identify all the candidate RSCs. To serve the purpose, the key feature sequences are extracted from an RSC to identify the bottleneck that have a great influence on the business process. This problem is called the key feature sequence of IIoT resource service chain (KFSR-IIoT). The proposed approach, composed of several algorithms, is called the Algorithms for KFSR-IIoT (AKFSR-IIoT). First, resource service modeling is introduced to obtain all the RSCs associated with a workflow. Next, the degree of influence (DI), which is the extent to which a feature of an upstream resource service influences a feature of a downstream one in an RSC, is analyzed and defined. After normalization, the average DI of any two adjoining features is calculated; thus, the key feature sequences are resolved. In the proposed algorithms, the average DIs have the characteristics of a high value, low variation, and long duration. Then, based on the DIs of key feature sequences, an algorithm is presented to measure the synthetic DI of an RSC. Finally, the AKFSR-IIoT is tested with a nearly practical data set, and the results show that it is very promising.