Research

Background and Motivation

The German government has developed a high-tech strategy called “Industry 4.0”, which shifts the industrial technology paradigm by adopting the key technologies: cyber-physical-systems, internet of things, services and people as the basis. This shift aims to interconnect human and the cyber-physical systems (i.e., physical objects, hardware, software), and also to allow them to communicate via an internet of things or services to collaborate and exchange data. However, these cyber-physical-systems are generally developed by different vendors. Therefore, they might have various structures, formats, and vocabularies for exchanging the data, and since the human stakeholders might have different knowledge background, the interoperability problems among the systems and these stakeholders might arise.

Each node in the internet of things generates data and is accessible via the internet. In our world, it can be imagined that there will be billions of such nodes. However, not all data provided by those nodes are useful to solve our problem. Hence, it could cause new problems that we are drowning in data, but starving for knowledge or useful information. Those big data are coming from heterogeneous sources. Ensuring the quality of those data, whether they are in a consistent form and able to fulfill the requirements of the applications using them, will be another challenge.

 

Research Focuses

Ontologies and Knowledge Graph

This approach employs a knowledge/information model as the integration element of heterogeneous schema and data sources. A W3C standard, called The OWL (Web Ontology Language), is used as the model representation which serves as a shared vocabulary and schema among the heterogeneous systems and humans [1]. Ontology allows conceptualization and formalization of human knowledge elements thus it creates an inter human-machine understandable model. OWL ontologies provide flexibilities to express logical statements through the ability to depict description logic and the possible integration of rules [2]. Therefore, logical reasonings, including fuzzy reasoning [19] are also possible in the knowledge model [3-4]. Different databases, which are annotated using ontologies, form linked data. It allows a flexible and efficient query/search of information from various domains that is already semantically linked.  Linked data will create a knowledge graph that will enable flexible structuring and representation of search/query results.

Semantic Uplifting and Linking

It aims to extract the semantic model of existing industrial data exchange schemas and formats, such as IFC, AutomationML, and CAD, in order to align them with the schema of the developed and reused ontologies.  The technologies that can be useful for this approach are for example ifc2owl and OntoCAD [5], which I have also involved in the community and their development.

In many use cases, the single industrial data exchange standard cannot fully fulfill the data requirements of an application. Meanwhile, there are growing initiatives and recommendations on sharing data and schemas from different domains on the internet (open data). I developed methods of linked data to reuse and interlink available data and schemas that are able to fulfill data requirements of an application driven by use cases [6-8]. The method shifts the approach to address interoperability issues from data exchange to data sharing.

Machine Learning

Data mining/machine learning techniques, such as regression, association analysis, classification, artificial neural network, and clustering, are artificial intelligence approaches to identify interesting patterns and dependencies in a collection of data. The information or knowledge required to solve certain problems are extracted, converted into ontology elements or rules [9], integrated with formalized human knowledge, and then stored in the ontology represented knowledge graph [10]. The rules gained from machine learning does not have 100% validity. Therefore, the rules are transformed into fuzzy rules that allows fuzzy assertion [19]. My research develops machine learning based techniques to solve problems in engineering and business, for instance, to predict the energy consumption of processes, to detect the energy wasting, to give recommendations for energy saving, to give recommendations for optimal factory and product configuration, etc. [1] [11] [12].

Explainable AI

The Explainable AI links the machine learning models with to the knowledge graph as an explainer. The nodes of the knowledge graph are used to annotate and describe the input, output, and hidden layers in the machine learning models. The knowledge graph explains the relationships between inputs, outputs, and the training data. Thus, it will bring the machine learning models to the right level of semantics and interoperability [11] [18].

Data-driven Simulation and Optimization

 The data-drive simulation and optimization approach has been used to improve production planning and scheduling. The existing exact scheduling approaches cannot give solutions for large problems within reasonable periods of time. Meta-heuristics, by contrast, work on general models that do not correspond to reality.  For this reason, a hyper-heuristic scheduling approach is developed yielding a flexible, but still real problem-relevant model [16]. The hyper-heuristic approach allows for the incorporation of different user-defined heuristics strategies, optimization objectives, constraints, and other configurations that come from ontology model. The model is also generated or adjusted using machine learning algorithm. The optimization objectives and constraints consider both economic aspects, e.g. lowest costs and avoiding certain shift work, and energy efficiency, e.g. lowest energy consumption, and avoiding peak power load [17].

Semantic Enrichment of Geometry Data to Improve Visualization and Interaction in VR/AR environment

 The OntoCAD tool extracts the semantic information from CAD drawings in a semi-automatic way. The drawing primitives from CAD files are used to perform the pattern matching and classification algorithms to extract the semantic information [5]. The resulting semantic information is then mapped to the corresponding ontology classes of a T-Box ontology. Finally, individuals of the corresponding classes are created to populate the ontology and their geometric properties like world coordinate position and bounding box are set. This approach enables the linking or annotation of geometry data with semantic information. By using the semantic information, the relationships between objects for visualization can be easily modeled. The complex interaction rules can also be modeled using SWRL or ontology axioms [2].

Applications of the Approach

The approaches described above have been applied in different domains, for example to make product configuration match customers’ needs through ontology based recommendation system [1] in the DIALOG[1] project, to improve resource and energy efficiency in buildings [2], [4-7], [10-11] in the KEHL[2], KnoHolEM[3], SERUM[4], and SWIMing[5] projects, to improve resource and energy efficiency in production [3] [9] [12] in WertProNET[6], wEnPro[7] and ecoBalance[8] projects, and to improve energy efficiency and to create business model in smart city context [8] [13] in DAREED[9] project.

I developed KPI models for energy and resource assessments, which are represented with ontology, since they involve different entities and processes within an organization that might have interoperability problems, e.g. machines, human stakeholders, buildings, organization units, and diverse influencing variables [14-15]. This allows optimization of energy and resource usage with the objective function to maximize the KPIs. I developed hyper-heuristic and meta-heuristic optimization methods to optimize the resource and energy consumption upon the ontology knowledge graph [16-17]. Thus, the optimization model contains both formalized human knowledge and knowledge extracted from data. The approach has been validated in several SMEs [18].

Applications in Digital Transformation for Sustainability

The application of the INDEED research approach focuses on a demand response system that engages power consumers, especially manufacturing, through price and incentive-based programs. The system is based on a data platform that collects data from heterogeneous systems at both supply and demand sides, which are linked through a semantic middleware, instead of centralized data integration. An ontology is used as the integration information model of the semantic middleware. The ontology is constructed using mapping techniques, reusing existing ontologies such as OpenADR, SSN, SAREF, etc., and applying ontology alignment methods. Machine learning approaches are developed to forecast both the power generated from renewable energy sources and the power demanded by manufacturing consumers based on their processes. The forecasts are the groundworks to calculate the dynamic electricity price introduced for the DR program. Different neural network architectures such as Deep Neural Network (DNN), Long Short-Term Memory Network (LSTM), Convolutional Neural Network (CNN), and Hybrid architectures are developed and linked to the ontology resulting in XAI. The energy consumption of the production processes is then optimized using hyper heuristic algorithms. The optimized energy consumption is recorded and transferred back to the energy provider. The whole system is developed as cloud-edge solution [23-25]

[1] http://www.dialog-das-projekt.de/

[2] http://kehl.actihome.de/

[3] http://www.knoholem.eu/

[4] http://www.imi.kit.edu/english/21_2626.php

[5] http://swiming-project.eu/

[6] http://www.wertpronet.de/

[7] http://wenpro.de/

[8] http://www.imi.kit.edu/21_2273.php

[9] http://www.dareed.eu/

 

[1] Wicaksono, H.; Schubert, V.; Rogalski, S.; Laydi, Y. A.; Ovtcharova, J. (2011): Ontology-driven Requirements Elicitation in Product Configuration Systems. In: Enabling Manufacturing Competitiveness and Economic Sustainability, ElMaraghy, H. A. (Ed.), Springer Verlag, Berlin, Heidelberg, pp. 63-67
[2] Wicaksono, H; Dobreva,  P.;  Häfner, P.; Rogalski, S. (2015): Methodology to Develop Ontological Building Information Model for Energy Management System in Building Operational Phase, Knowledge Discovery, Knowledge Engineering and Knowledge Management Communications in Computer and Information Science, Springer, Volume 454, 2015, pp. 168-181
[3] Wicaksono, H.; Rogalski, S.; Jost, F.; Ovtcharova, J. (2014): Energy Efficiency Evaluation and Optimization in Manufacturing through Ontology Represented Knowledge Base, in: International Journal Intelligent Systems in Accounting, Finance and Management, John Wiley and Sons, 2014, pp. 1099-1174
[4] Wicaksono, H.; Dobreva, P.; Häfner, P.; Rogalski, S. (2013):  Ontology Development towards Expressive and Reasoning-enabled Building Information Model for an Intelligent Energy Management System, 5th International Conference Knowledge Engineering and Ontology Development 2013, Vilamoura, Algarve, Portugal, 19-22 September 2013
[5] Häfner, P.; Häfner, V.; Wicaksono, H.; Ovtcharova, J. (2013): Semi-automated Ontology Population from Building Construction Drawings, 5th International Conference Knowledge Engineering and Ontology Development 2013, Vilamoura, Algarve, Portugal, 19-22 September 2013
[6] Wicaksono, H. (2017): Einsatz von (Open)-Linked-Data für Gebäudeinformationsmodellierung, Digitale Transformation des Baubetriebs, Best Practice und Zukunftstrends, Forum auf der Messe BAU 2017, Munich, 19 January 2017
[7] McGlinn, K.; Weise, M.; Wicaksono, H. (2016): Towards a Shared Use Case Repository to support Building Information Modelling in the Energy Efficient Building Domain, CIB W78 Conference, Brisbane, Australia, 31 Oct – 2 Nov 2016
[8] Tonev, K.; Wicaksono, H. (2016): Semantic data integration for smart cities using linked data, 11th European Conference on Product and Process Modelling, Limassol, Cyprus, 7-9 September 2016
[9] Wicaksono, H.; Rogalski, S.; Ovtcharova, J. (2012): Knowledge Management Approach to improve Energy Efficiency in Small Medium Enterprises, 10th International Conference on Manufacturing Research (ICMR 2012), 11-13 September 2012, Birmingham, UK
[10] Wicaksono, H.; Aleksandrov, K.; Rogalski, S. (2012): An Intelligent System for Improving Energy Efficiency in Building Using Ontology and Building Automation Systems, in: Kongoli, F. (Ed.), Automation, Publisher: InTech, Open Access Publisher, pp. 531-548
[11] Wicaksono, H.; Rogalski, S., Kusnady, E. (2010): Knowledge-based Intelligent Energy Management Using Building Automation System, Proceeding 9th International Conference on Power and Energy, 27-29 October 2010, Singapore
[12] Wicaksono, H.; Ovtcharova, J. (2012): Energy Consumption Regulation and Optimization in Discrete Manufacturing through Semi-automatic Knowledge Generation using Data Mining, International, 10th Global Conference of Sustainable Manufacturing (GCSM), 31 October – 2 November 2012, Istanbul, Turkey
[13] Wicaksono, H.,Tonev, K. (2016): Linked Data in the DAREED Smart City project, 4th International Workshop on Linked Data in Architecture and Construction (LDAC), Madrid, Spain, 21-22 June 2016
[14] Rogalski, S.; Wicaksono, H. (2011): Methodology for Flexibility Measurement in Semi-automatic Production.  In: Enabling Manufacturing Competitiveness and Economic Sustainability, ElMaraghy, H. A. (Ed.), Springer Verlag, Berlin, Heidelberg, pp.141-146
[15] Wicaksono, H.; Belzner, T.; Ovtcharova, J. (2013): Efficient Energy Performance Indicators for Different Level of Production Organizations in Manufacturing Companies, APMS 2013 International Conference Advance in Production Management Systems, State College, PA., 9-12 September 2013
[16] Wicaksono, H.; Prohl, E.V.; Ovtcharova, J. (2013): Hyper heuristc based production process scheduling to improve productivity in sustainable manufacturing, the 22nd International Conference on Production Research, Brazil, 28 July – 1 August, 2013
[17] Wicaksono, H.; Aleksandrov, K.; Rogalski, S.; Ovtcharova, J. (2013): An ICT Supported Holistic Approach for Qualitative and Quantitative Energy Efficiency Evaluation in Manufacturing Company, Competitive Manufacturing for Innovative Products and Services, IFIP Advances in Information and Communication Technology, Springer
[18] Wicaksono, H. (2016): An Integrated Method for Information and Communication Technology (ICT) Supported Energy Efficiency Evaluation and Optimization in Manufacturing: Knowledge-based Approach and Energy Performance Indicators (EnPI) to Support Evaluation and Optimization of Energy Efficiency, Dissertation, Karlsruher Institut für Technologie (KIT)
[19] Howell, S.; Wicaksono, H.; Yuce, B.; McGlinn, K.; Rezgui, Y (2018): User Centered Neuro-Fuzzy Energy Management Through Semantic-based Optimization, IEEE Transactions on Cybernetics doi: 10.1109/TCYB.2018.2839700
[20] Haas, K.; Kappe, S.; Siebert, M.; Wicaksono, H.; Ovtcharova, J. (2018):  Digital Assistance Based on an Ontology Driven Model of the IT-Systems Along the Product Lifecycle, Debruyne, Christophe; Panetto, Hervé; Weichhart, Georg; Bollen, Peter; Ciuciu, Ioana; Vidal, Maria-Esther; Meersman, Robert (Ed.): Springer, 2018, ISBN: 978-3-319-73805-5.
[21] Tonev, K.; Kappe, S.; Krahtova, P.; Wicaksono, H.; Ovtcharova, J. (2018): District-Scale Data Integration by Leveraging Semantic Web Technologies: a Case in Smart Cities, Christophe Debruyne Hervé Panetto, Georg Weichhart Peter Bollen Ioana Ciuciu Maria-Esther Vidal Robert Meersman (Ed.): pp. 289-292, Springer, 2018, ISBN: 9783319738055.
[22] Schneider G.F.; Wicaksono, H.; Ovtcharova, J. (2019): Virtual Engineering of Cyber-Physical Automation Systems: The Case of Control Logic, Advanced Engineering Informatics, Volume 39, pp. 127-143
[23] Wicaksono, H.; Yuce, B.; McGlinn, K. (2021): Smart Cities and Buildings, in Buildings and Semantics: Data Models and Web Technologies for the Built Environment, Taylor and Francis (Forthcoming)
[24] Wicaksono, H.; Boroukhian, T.; Bashyal, A. (2021): A Demand-Response System for Sustainable Manufacturing Using Linked Data and Machine Learning, Dynamics in Logistics, Springer (Forthcoming)
[25] Accelerating energy transition to green electricity through artificial intelligence, ICIMECE 2021, Surakarta, Indonesia, 5 October 2021, https://doi.org/10.31219/osf.io/tcrkh

Current Project

Talenta

BMDV, 01/2023 – 06/2025

In the transportation sector, the implementation of digital twins is part of the digitization measure to improve resource efficiency in infrastructure management. However, the use of digital twins is still limited due to challenges such as lack of common understanding of digital twin models, difficult model integration, security issues, lack of access to important data, and high costs due to inefficient business models.

The aim of the project is the development of an asset management platform suitable for SMEs for the cross-company, secure and intuitive collaborative management of assets of the digital twins. This can be achieved by developing a standardized graph-based semantic model of the asset, explainable machine learning (XAI) and a scenario-based intelligent search and discovery mechanism of the asset.

As part of the project, a uniform semantic description of digital assets using ontology is being developed in order to promote a common understanding of data and models between companies/organizations. Data sources with different formats are linked to the ontology by AI approaches and led to a knowledge graph. The project is developing an XAI toolset/model library to improve the transparency of the information obtained from the digital twins.

Delfine (Digitalization of Energy Transition)

BMWK, 08/2020 – 07/2023

Delfine aims to accelerate the adoption of demand response systems. It is an interdisciplinary project and develops a solution for the participation of industrial end customers in both price and incentive-based DR programs. The aim is to determine the influence of such programs on the network as such and on the development of electricity costs in the manufacturing industry. With the support of Stadtwerke Trier as a network operator and an interdisciplinary consortium with complementary competencies, this project strives for a technical solution that can be used in various areas. A continuous data network is developed using semantic middleware, from the automated creation of generation and demand forecasts to the dynamic design of electricity prices and the energy-efficient and intelligent use of production resources. The holistic consideration of the addressed issues by the interdisciplinary project consortium enables business models to be developed for the use of the project results by electricity providers, aggregators, and the manufacturing industry.

Delivery Assurance & Operational Supply Chain: Enhancing delivery performance with decision targeted analytics using causal machine learning

Industry project, 11/2021 – 10/2024

Supply Chains (SC) are complex, data driven systems which deal with the flow of information, goods, services, and money. In such fast-paced environments, the lack of data has been replaced by concerns of abundance of data, which has created difficulties in properly synthesizing the data and deriving meaningful conclusions from it. This makes it challenging to establish the relationships between different components of the SC. Nowadays, ‘optimized’, ‘resilient’ and ‘fail-safe’ are key words for a successful SC, however, conventional Machine Learning (ML) approaches have proven to lack the capability to provide the necessary tools to support SCs in fully achieving their goals. ML has found its way in many use cases in SCM such as predictive demand forecasting, intelligent partner selection, and assistance systems for resource management. However, the heavily statistical mode of most ML systems entails several limits on their power and performance. In a world filled with uncertainties (political, social, environmental etc.) it can be difficult to establish a resilient SC, especially without fully understanding the cause effect relationships between its different external and internal KPIs. Establishing and understanding such relationships would aid in a more effective Risk Management system for SCs, by minimizing disruptions when there are challenges in any stage of the SC.

Current ML approaches are not fit to optimize the movement of people or goods around the globe, as they are insufficient for robust predictions and reliable decision-making based on their correlational pattern recognition nature (causaLens). That is because current ML models are based on past patterns that may not hold in the future. They also produce hard-to-trust black box predictions without fully understanding the business context, therefore, making predictions rather than recommendations. Businesses ultimately want to do more than just predict the future; they also want to actively shape it by fully understanding and utilizing their data. Causal Machine Learning (CML), an emerging field in AI, could prove to be a very helpful tool in adding the missing links to SC, since identifying causal effects is an integral part of understanding and learning from behaviors, as well as shaping them. The proposed doctoral research project aims to address these challenges. The research project utilizes multi criteria decision making methods for interviewing experts to first identify and quantify the external risk factors, internal supplier performance KPI’s and find causal relationships. All the knowledge of experts will be then used in Causal Machine Learning Model to help have a better-informed decision and business recommendations.

The research defines three milestones to be achieved. Each objective has research questions to be answered by the research. Those research objectives are Quantification and utilizing of Experts Experience, to make the Operational Supply Chain more proactive by early prediction of escalated suppliers and their KPI’s and to know the Effect of External Risk Factors and other KPI’s on the suppliers and to give business recommendations for the next steps that helps reduce human biases and increase the proactiveness.

 

Past Projects

SWIMing (Semantic Web for information modeling in energy efficient building)

EU Horizon 2020, 02/2015 – 01/2017

The ‘SWIMing’ project brings together existing EeB projects under clusters categorised by which stages of the BLC the project is applied and energy savings are achieved and the particular domains within those stages to facilitate knowledge sharing and increase the impact of project results.

It supports making the project processes and the data produced by these processes publishable as linked data on the web and through this structured format (BIM-LOD), make data more accessible and less fragmented across the BLC. This data sharing will address the requirements that span application domains and BLC stages related to energy efficient practices in buildings, thus easing the exploitation of project results beyond the specific application evaluated in individual projects. The approach will also align projects with the wider knowledge-driven economy made possible by the open technology and broadening skill base growing around open web data standards and the resulting technologies will help reduce energy use and Greenhouse Gas (GHG) emissions worldwide, by making Europe a competitive leader in the green building industries.

DAREED (Decision support advisor for innovative business models and user engagement for smart energy efficient disctrics)

EU FP7, 09/2013 – 12/2016

DAREED aims at delivering an IT service oriented platform to support decision-making and foster energy efficiency at district level. Project results will be validated via five pilots from three EU countries and, thus granting the possibility to generalize results and ensuring sustainability throughout Europe. The project targets decision-makers and energy providers at local level and is seeking to support them to overcome technological, financial and knowledge barriers and adopt new strategies for improving the energy performance.

KlimaRa

The aim of the projects is to develop a “KlimaRa” portal solution for user-dependent, knowledge-based energy consumption control in selected rooms of the town halls. A necessary prerequisite is the additional equipment of the premises with special sensors and actuators. This is necessary, on the one hand, to record energy data in a consumer-specific manner and, on the other hand, to avoid major renovation work in the listed building by using Powerline as the in-house data connection. When selecting the room, care was taken to ensure that the rooms were connected to a circuit / line and on the same floor, if possible, in order to keep the installation effort low.

On this basis, it is necessary to evaluate the different consumption patterns in the rooms of the town hall to be examined and to offer room users the opportunity to intervene in the consumption control system by means of intelligent, automated mechanisms in order to save energy by using mobile devices (e.g. smartphones or tablets). At the same time, the intention was to raise the awareness of various building users by displaying their own energy consumption with the coupled CO2 emissions individually and transparently.

 ecoBalance

BMWi/ZIM, 01/2012 – 07/2014

The aim of the project to develop an intelligent frequency converter that records the current module-related energy consumption and controls the power consumption through communication with other modules of one or more machine tools so that high individual load peaks are avoided. Machine and energy data are to be saved with product and machine properties in a knowledge base and energetic performance profiles for individual machining operations are to be determined. By coupling a so-called sequence calculation unit with the knowledge base, machine orders will in future be automatically evaluated and controlled in such a way that the power is distributed as evenly as possible without high load peaks, with an intelligent data monitor allowing manual control interventions and forecasting expected energy load peaks for new processing orders.

KnoHolEM (Knowledge-based energy management for public buildings through holistic information modeling and 3D visualization)

EU FP7, 09/2011 – 08/2014

The KnoholEM main objective is to improve energy efficiency in public buildings trough progressive and intelligent knowledge modeling approach able to respond to changes in building configuration and use.The project methodology takes into account and overcome challenges related to high variety of dynamic factors with complex interactions and integration of new standards (IFC). The main project innovations are related to an extended alignment with standards (IFC-OWL mapping, CAD/IFC to OWL), creation of an intelligent knowledge base balanced by several ontologies as heart of the system architecture; 3D visualization of the building environment for the final user and an automated ontology population concept that go behind the state of the art technologies of similar EU projects.

 

wEnPro

BMBF, 06/2011 – 05/2013

Development of methods for analyzing, evaluating and improving energy and resource efficiency in manufacturing SMEs. These are to be integrated into a modular toolset and provide, in addition to pure monitoring functions, in particular, knowledge-based evaluation and control functions. The toolset improves the sustainability of resources usage whilst maintaining flexibility in production.