Automating simulation: an open source software for automated input data in discrete event simulation projects

Discrete Event Simulation (DES) is a well-established decision support tool in modelling work flows in manufacturing industry. However, despite its potential, manufacturing industry has been less than successful in the implementation of simulation as a decision support tool. One of the main weaknesses of operating DES is the extensive time-consumption of simulation studies. The reason for the large time-consumption is arguably due to the input data process. The time spent on the input data process is on average 32% of the total time of a DES project (Onggo, Hill et al. 2013). The main purpose of this thesis is to enable the constant use of simulation by various manufacturing personnel as a decision support tool. The aim is to reduce the time-consumption for the input data process, and thus facilitate the provision for automated update of operational data to DES models. The thesis starts with a literature review on the automation of input data to DES. This work helped us understand the nature of the input data phase. By examining the proposed methodologies and tools, we established that automation of input data is required to support real-time simulation. Next, a survey of OS data science tools is presented. The main purpose of this survey was to identify the most appropriate tool that would be the basis for the development of the KE tool. The outcome of this survey led to the selection of Python and the RPy2 (http://rpy.sourceforge.net) library as the base for the KE tool. In the next part of the thesis we describe the development of the KE tool. The open source KE (“Knowledge Extraction”) tool consists of features, written in the Python programming language, that enable the three main components: Data extraction; Data processing; Output preparation. The tool reads data from several organisations’ resources; analyses it and outputs it in a format that is applicable to be used by a simulation tool, all conducted in one automated process. The KE tool aims to offer an open source automated solution that reduces the time-consumption in the input data process in DES projects. The KE tool is licensed under the terms of GNU Lesser General Public Licence (LGPL), the code of the tool is released in GitHub (https://github.com/nexedi/dream/). In the repository one can find and download the developed objects themselves, examples with the development of the KE tool in different simulation topologies and a thorough documentation of the tool. The code is kept under version control with Git (http://git-scm.com/); the user can clone and manipulate the different versions of the code through the project repository in GitHub. Then, a description of the deployment of the KE tool in a large scale organisation with availability in data resources is presented. This helped us validate the tool in a data rich company for usability in real world situations. Based on the validation results, the use of the KE tool reduced the needed time for arranging input simulation data from 350 minutes (5 hours and 50 minutes) to 65 minutes (1 hour and 5 minutes), a reduction of 81% contrasted to the traditional manual approach. Next, in our efforts to prove the versatility of the tool and after identifying that the role of DES within Small and Medium Enterprises (SMEs) has in general, received little attention by academics and simulation practitioners, we developed an approach that facilitates and supports the use of real-time DES in SMEs. The approach is designed considering the financial limitations of SMEs; therefore it doesn’t require the investment of money on acquiring software applications. The SME approach that utilises the KE tool for the supply of real-time data to the simulation model was tested in two pilot cases. We conclude the thesis summarizing with our findings and future research topics.