Research On Discovery Method Of Structure And Role Type Based On Communication Behaviors In Complex Collaborative Work System

With the development of computer science and information technology, systems with complex collaboration have been found in many areas, and structural design of these systems is constantly improved and optimized in the long-term evolution process. However, many change information in the improve process is often the lack of corresponding change records, resulting in associated systems management and researchers hard to grasp the local detail information and the overall structure information quickly, accurately and comprehensively. But these information are very important for optimized design of system architecture, fault localization and fault propagation path detection and so on.In this paper, we try to solve this problem by designing a discovery algorithm based on the communication behaviors between communication individuals in system. Since the communication devices in the system tend to have implicit hierarchy, this paper proposes a discovery algorithm applicable to discover the organization structure and the role type of communication individual in the complex cooperative work systems with three-layer hierarchy.Firstly, we get peer communication relationship between devices through a comprehensive analysis of the multicast communication message in the system and the log information on the individual in and out message group, and we build a communication network between communication individuals based on these relation information; then, we construct the attribute set of the communication individuals and get the hierarchical relationship between the individuals by using sub-algorithm on dividing hierarchy relationship based on the communications network; next, we get the role information of devices in the communication process by using sub-algorithm on dividing role based on the hierarchical relationship between the communication individual; Finally, we get the organization structure inside the system by using infomap algorithm and sub-algorithm on dividing organization structure based on the hierarchical relationship between devices and the role information of individuals.We carried out experiments on the simulation dataset based on the communication behavior in a cooperative system in a research laboratory. At the same time, we compared the result with the infomap algorithm. According to the organization structure given by experts of the military unit, we evaluated the experimental results by using the NMI evaluation index of community finding. Evaluation results show that the results of our experiment are better than the results of infomap algorithm, it also reflects the rationality and effectiveness of the method proposed in this paper.