Research On Mining Methods And Analysis Technologies Of Business Process Model And Organization Structure Model

For adapting to constantly changeable market environment, enterprise manage way is changing from trational function management to business process management (BPM). And applying of business process management system (BPMS) becomes a realization instrument. With the rapid development of information technology, automatic drive technology is not a bottle-neck to affect the capability of BPMSs. The ability of BPM instead becomes the main factor.Business processes are the base of enterprise running. At the same time appropriate enterprise organization structure is the necessary support to realize process management. With the continuous change of the market, business process and organization structure are continually changed. Therefore, the modeling methods and analysis technologies of dynamic processes and organization structure are necessary. Because there are much shortage existed in current modeling methods and analysis technology, this paper researches novel modeling methods based on process logs and designs corresponding rationality analysis technologies in order to improve the ability of BPM.Firstly, with the accumulation of process logs in BPMS, how to use the practical data to find the change of process is the front of BPM. The new method is called as process mining. This method is based on data mining and primarily satisfies the modeling demand of dynamic process. Currently the method is in its infancy, all algorithms describe process model through single formalization method. Therefore the performance of algorithm is affected by special formalization expression method so that its generalization is worse. Thereby, this paper researches a new process mining method which independly designs process mining algorithm and formalization expression algorithm. The method analyzes the transfer probabilities among the tasks and constructs a transfer matrix. Based on the matrix and predefined logic relation identification rules a structure mining algorithm is designed. Based on the mining result formalization expresses algorithms based on directed net graph and WF-net are designed.Secondly, because verification methods of rationality of process structure in current qualitative analysis technologies are scarce, this paper researches a rational analysis technology based on multi process modeling methods. At first, a group of mapping rules is designed, which transform the process from directed network graph to WF-net. Next, a verification method of WF-net's rationality based on simplifying rules is researched, which not only can automatically verifies the rationality of WF-net, but can place the abnormal structures.Current quantitative analysis technologies are limited in computation of key performance indicators (KPIs) and neglect the abnormal situation that process instances deviats from predefined model so that the validity of process analysis is affected. Thereby, this paper designes a business process quantitative analysis technology by computing and analyzing process deviation degree based on sequence alignment technology. At first, this paper designes a framework of process validation. Next, four methods to computer deviation degree based on different sequence distances are designed. At last process analysis rules based on deviation degree are provided.Thirdly, because current organization structure modeling methods are difficult to describe dynamic structure, this paper reaserch on a team-based organization structure modeling method based on process logs, which are called as organization structure mining. In this method team relation network graph, role relation network graph and authorization relation network graph are used as the dynamic team-based organization structure describing model. Furthermore this paper defines relation mining rules and measurement metrics by analyzing different relations among members based on logs. And based on these rules and metrics the modeling methods of relation network graphs are designed.Fourthly, because of shortage of analysis technology in organization structure model, this paper research on a rationality analysis technology of team-based organization from the view of relation based on social network analysis technology. At first, analysis of team-based organization structure model is simplified to be analysis of corresponding team relation network graph. Next, two rationality analysis technologies based on centricity and cohesive subgroups are designed, which verify rationality of network graph by computing the power and analyzing distribution of groups.Finally, for verifying realizableness of methods and technologies, based on above research, this paper designs a mining and analysis prototype system of business process model and organization structure model. This paper concretely designs the functions of evry modules and data transfer relations among modules based on system demand analysis.This thesis solves some of the critical problems and deficiencies existing in dynamic process modeling, organization structure modeling, and corresponding model analysis technologies at a certain extent, which can provide a technology support to enhance the capability of BPM.