ตัวจำลองการทำงานในสำนักงานบนระบบจัดการกระแสงานอูซซี

Abstract

This thesis presents a simulation of automatic workflow management system to support any office or organization expansion. The goal of this research is the scalability of workflow simulation which can be utilized even the system of organization is scaled. Distributed processing system is used for both data collection and workflow execution of the simulation system. The number of person and work roles is considered to be applied in this proposed workflow simulator for the purpose of execution and processing workflow to make the simulation more accorded to business conditions. This research focused on decreasing of delay time which is waiting time of each person before working. Therefore, the overall process of workflow has less delay time and succeed. We proposed the simulation system to control the workflow simulation processing. The main functions are start, stop, pause, log, and export data. User or administrator controls the simulator via the web user interface. Data storing is designed and implemented to support big data processing. Therefore, the Apache HBase was adopted to store the workflows and workers data for the flexibility and scalability. The Apache HBase is a database which has non-relation of storing data architecture. Furthermore, we considered to adopt both the Apache Oozie and the Apache Hadoop which are the workflow engine for workflow simulation execution on distributed sys- tem. We tested by running 100 workflow simulation. The results showed that it causes the delay time in some of the work roles in workflow process. In addition, we added 2 peoples into the work roles which has maximum delay time, and then retesting. The retesting result showed delay time is 0%.