Research On Optimization Method Of Organization Model For Business Process

As the environment continues to change,enterprises need to constantly adjust their organizational structure to respond to various personnel transfers and other dynamic demands.Having a flexible and appropriate structure in human resources can help address such challenges,so organizations need to acquire and maintain an accurate and timely understanding of human resource groupings.Existing static organizational architecture diagrams cannot meet this requirement,but organizational structure mining can help to meet this requirement by using data related to business processes to discover resource groups with similar characteristics during process execution.However,the existing organizational structure mining will still face two major challenges :(1)to evaluate the organization model accurately in many aspects.(2)Optimize the evaluated organizational model.This paper focuses on the research of organizational structure mining.Aiming at the above challenges,this paper improves and extends the existing organizational model evaluation methods,and proposes an organizational model optimization method based on a genetic algorithm.Finally,an organization model mining and optimization tool are designed and implemented.The main content of this paper is as follows:Firstly,the existing evaluation methods of organizational models are improved and extended.Aiming at the problem that fitness does not punish the logging behavior,the concept of resource penalty is introduced to improve fitness.Aiming at the problem that the accuracy does not consider the disallowed events,the accuracy is improved by introducing the influence of disallowed events on the organization model.Aiming at the problem that the accuracy is not suitable for multi-candidate resource scenarios,the model evaluation is carried out by considering the difference in the execution modes of resources in the model and log,and the similarity of resource execution modes is proposed.Aiming at the problem that internal evaluation of the organizational model cannot be carried out,an improved contour coefficient is proposed by referring to the internal evaluation method of clustering-contour coefficient.Based on the above four evaluation methods,the purpose of a multifaceted evaluation of the organization model is realized.In the second.This paper presents a method of organization model optimization based on a genetic algorithm.When the mined organizational model is not good enough,further optimization of the organizational model can be considered.By expressing the results of the execution pattern in the organization model in the form of a matrix,the genetic optimization method is used to obtain a better matrix,so as to obtain better organization model.In the genetic optimization algorithm,the matrix obtained from the initial population is taken as the population individual by the existing execution mode allocation methods: full allocation method and whole analysis method.The population fitness is calculated by the direct transformation method,and the evaluation result is taken as the population fitness.Moreover,the elite retention strategy is adopted to ensure that the best individuals are not lost.On the selection operator,the better the individual,the easier the roulette selection method is adopted.For the crossover operator,a uniform crossover method was used in which each gene of subclass was randomly determined by one parent.The mutation operator adopts a multi-bit inversion mutation method which selects multiple genes to flip over and introduces a mutation probability matrix which can improve the search efficiency.The experimental results show that the genetic optimization method proposed in this paper can indeed play the optimization effect,and can obtain a better tissue model.Thirdly,in order to verify the usability of the proposed method,this paper designs and develops an organization model mining and optimization tool.The overall design,outline design,and function design of the tool are introduced successively,and finally,the function is realized.With this tool,the purpose of obtaining,evaluating,and visualizing organizational models from event logs can be realized.