Monostori, L.; Ilie-Zudor, E.; Kádár, B.; Karnok, D.; Pfeiffer, A.; Kemény, Zs.; Szathmári, M.: Novel IT solutions for increasing transparency in production and in supply chains, Proceedings of the 21st International Conference on Production Research, ICPR 21, July 31 – August 4, 2011, Stuttgart, Germany, (ISBN: 978-3-8396-0293-5) paper no.: IS1.3, p. 6. NOVEL IT SOLUTIONS FOR INCREASING TRANSPARENCY IN PRODUCTION AND IN SUPPLY CHAINS L. Monostori 1,2 , E. Ilie-Zudor 1 , B. Kádár 1 , D. Karnok 1,2 , A. Pfeiffer 1 , Zs. Kemény 1 , M. Szathmári 1 1 Fraunhofer Project Center on Production Management and Informatics, Computer and Automation Research Institute, Budapest, Hungary 2 Department of Manufacturing Science and Technology, Budapest University of Technology and Economics, Budapest, Hungary Abstract The paper summarises the main challenges and problems related to transparency in production firms and networks. The transparency problem of production are treated from three aspects, i.e., factory level data gathering tight-coupled with productions simulation; extracting knowledge from large, complex, time- dependent noisy and anomalous process logs; and identity-based tracking and tracing services within and beyond organizational borders. For all these fields of industrial relevance novel approaches and solutions are presented. Keywords: Transparency, Production; Information and communication technologies, Data mining, Auto identification 1 INTRODUCTION The real-time management of changes and disturbances in production execution along with the distributed effort of supply chain planning and management have become significant in technical, economic and operational aspects of today's production companies. The practical feasibility of most of the approaches to production modeling and control boils down to providing sufficient information about the involved processes and entities [1]. In order to master the high dynamics in the processes and demand, real-time feedback from production is required. For real-time control actions, information about the state of the controlled system must be provided without large time lags. While information flow is easier to manage within and between IT components, it may become critical to maintain links between physical products and the related software agents, as the product is continually changing and moving without a permanent network connection being guaranteed [2]. Better information flow and transparency may also contribute to further improvement, such as real event-driven control [3], as well as ‘plug and produce’ performance [4] based on autonomous resources and intelligent products. The paper addresses the transparency problem from three aspects. Firstly, data gathering on factory level is focused on, with special emphasis upon the mirroring of the current state of the production into the digital world in order to support production planning and control decisions (Section 2). Unfortunately, the huge amount of data collected during production needs data filtering and mining, because of missing and / or distorted information. Novel approaches to this issue are presented in Section 3. Section 4 concentrates on tracking of products and product data within and beyond organizational borders by introducing an open-source solution platform for this purpose. 2 DATA GATHERING AND SIMULATION AT FACTORY LEVEL One of the utmost requirements regarding transparency in manufacturing systems at factory level is the availability of actual, reliable productions systems’ status data transformed and provided as information according to the users’ requirements. In a large-scale manufacturing environment taking the most appropriate control decision, as well as the prediction of waiting times, workloads or utilisation of the resources is a difficult task and hardly depends on the availability of data. These control decisions and the analysis of their effects can be supported by simulation- based analysis relying on proper data acquisition and data transformation methods [5], [6]. A discrete-event simulation-based analysis system is proposed in this section which supports both the automatic mirroring of the current production system’s state in a digital model, and the analysing and validating of different production control settings and rules in the digital model. Due to the large number of the resources, the relatively long and frequently reengineered routings and the often changing product types, the maintenance of the simulation model and the provision of up-to-date input data are always difficult tasks [7]. Thus, one of the limitations of using simulation in the on-line decision making process is the considerable amount of time spent on gathering and analysing data. In quasi real-time control (hours, minutes), however, the three key issues are. • Data acquisition and validation for simulation input. • Quick response time of simulation runs and analysis. • The ability of creating the snapshot of the physical system status in the simulation model by instantaneous feedback. The main goal of the research and development presented here was to enhance the simulation-based analysis and dispatching system by eliminating the manual data acquisition through automatic interfaces, to create a more realistic model of the factory and to improve the dispatch logic of the control system. Furthermore, the self-building simulation model now provides prospective (e.g., locating anticipated disturbances, identifying the trends of designated performance measures), and retrospective (e.g., gathering statistics on resources) simulation functionalities.