Analysis And Modeling Of Organizational Network Based On Data Of Workflow Logs

There are various organizations in real world with different scale and different missions.The organization deals with exchange relations in a three-dimensional connection and collaborates to accomplish all the missions in it.Thereby forming a variety of autonomous units and task-oriented modules.The increasingly frequent interaction within or between these units or modules enables the organization to gradually present a complex network morphology and show the same behavioral patterns and characteristics with complex network which is common in social reality.Based on the premise of this discovery,this paper introduces the theory of complex network into the research of organizational network.Taking a real organization as an example,this paper uses web crawlers to capture data of workflow logs in the organization,and builds organizational networks based on it.Then we analyze the topological structure characteristics of organizational networks and design organizational network evolution models.Some basic problems in the research of organizational network research was discussed,providing factual basis and application background for the exploration of new theories.This study mainly carried out the following work:1)Introduced the functions and characteristics of OA system,summarized the workflow maps in OA system,and analyzed the classification and characteristics of workflows.The collection,preprocessing and storage of workflow logs data in the OA system were discussed,and the organization network data set based on real workflow logs was constructed.2)From the aspects of workflow management,creation and circulation time of workflows,this paper analyzed the behavior characteristics of the organization members in the OA system including time preference and processing efficiency.3)We constructed the massage network model and reimbursement network model based on workflow logs data and visualized them by applying complex network modeling method.On this basis,the topology structure characteristics of the organization network were analyzed from a static perspective,such as Clustering coefficient,average path length,Modularity,matching coefficient,degree distribution,edge weight distribution,etc.Dynamical topology characteristics of the organizational networks are studied base on the semester time interval.4)This paper studied the evolutionary characteristics of the message network and reimbursement network,and designed a weighted-undirected network evolution model and weighted-directed network evolution model.The simulation experiments were used to verify the consistency between the two evolution models and the actual organization networks.The study suggests that both the message network and the reimbursement network are assortative network,with small-world features and clear community structure features.Degree distribution,intensity distribution,and edge weight distribution don't obey the power law distribution.Strengths and degree of nodes do not have the power function association like s~k~?.The two network evolution models designed in this paper conform to the statistical and topological features of the organization networks.