@inproceedings{Brand.2024_LLMKnowledgeEngineer,
  author       = {Brand, Florian and Malburg, Lukas and Bergmann, Ralph},
  title        = {{Large Language Models as Knowledge Engineers}},
  booktitle    = {Proceedings of the Workshops at the 32nd International Conference
                  on Case-Based Reasoning {(ICCBR-WS} 2024) co-located with the 32nd
                  International Conference on Case-Based Reasoning {(ICCBR} 2024), M{\'{e}}rida,
                  Mexico, July 1, 2024},
  series       = {{CEUR} Workshop Proceedings},
  editor       = {Lukas Malburg},
  volume       = {3708},
  pages        = {3--18},
  publisher    = {CEUR-WS.org.},
  year         = {2024},
  abstract 	   = {{Many Artificial Intelligence (AI) systems require human-engineered knowledge at their core to reason about new problems based on this knowledge, with Case-Based Reasoning (CBR) being no exception. However, the acquisition of this knowledge is a time-consuming and laborious task for the domain experts that provide the needed knowledge. We propose an approach to help in the creation of this knowledge by leveraging Large Language Models (LLMs) in conjunction with existing knowledge to create the vocabulary and case base for a complex real-world domain. We find that LLMs are capable of generating knowledge, with results improving by using natural language and instructions. Furthermore, permissively licensed models like CodeLlama and Mixtral perform similar or better than closed state-of-the-art models like GPT-3.5 Turbo and GPT-4 Turbo.}},
  keywords 	   = {{Case-Based Reasoning, Knowledge Engineering, Knowledge Acquisition Bottleneck, Large Language Models, Prompting}},
  url = {https://www.wi2.uni-trier.de/shared/publications/2024_ICCBR-WS_LLMInCBR_BrandEtAl.pdf}
}

Many Artificial Intelligence (AI) systems require human-engineered knowledge at their core to reason about new problems based on this knowledge, with Case-Based Reasoning (CBR) being no exception. However, the acquisition of this knowledge is a time-consuming and laborious task for the domain experts that provide the needed knowledge. We propose an approach to help in the creation of this knowledge by leveraging Large Language Models (LLMs) in conjunction with existing knowledge to create the vocabulary and case base for a complex real-world domain. We find that LLMs are capable of generating knowledge, with results improving by using natural language and instructions. Furthermore, permissively licensed models like CodeLlama and Mixtral perform similar or better than closed state-of-the-art models like GPT-3.5 Turbo and GPT-4 Turbo.

@inproceedings{Schultheis.2024_MissingSensorValues,
  author       = {Schultheis, Alexander and Malburg, Lukas and Grüger, Joscha and Weich, Justin and Bertrand, Yannis and Bergmann, Ralph and Serral Asensio, Estefanía},
  title        = {{Identifying Missing Sensor Values in IoT Time Series Data: A Weight-Based Extension of Similarity Measures for Smart Manufacturing}},
  booktitle    = {Case-Based Reasoning Research and Development - 32nd International Conference, {ICCBR} 2024, Merida, Mexico, July 1-4, 2024, Proceedings},
  series       = {Lecture Notes in Computer Science},
  pages        = {240--257},
  volume       = {14775},
  publisher    = {Springer.},
  year         = {2024},
  doi          = {10.1007/978-3-031-63646-2\_16},
  abstract 	   = {{Smart Manufacturing integrates methods of Artificial Intelligence and the Internet of Things into processes to enhance efficiency and flexibility. However, analysis of time series sensor data, crucial for process optimization, is susceptible to Data Quality Issues (DQIs) and can lead to operational problems. Traditional machine learning approaches struggle with limited error data availability in addressing DQIs. The knowledge-driven approach of Case-Based Reasoning targets this issue by reusing experiences regarding already identified DQIs. While some DQIs can be detected using conventional similarity measures, the common, frequently occurring DQI type of Missing Sensor Values pose challenges that cannot be solved using established measures. To address this, this paper proposes a weight-based extension of similarity measures for time series data. This extension aims at the identification and handling of missing sensor values in smart manufacturing processes. Furthermore, analog extensions of established time series measures are presented and possible areas of application outside the DQI domain are outlined.}},
  keywords 	   = {{Temporal Case-Based Reasoning, Time Series Data, Time Series Similarity Measures, Data Quality Issues}},
  url = {https://www.wi2.uni-trier.de/shared/publications/2024_ICCBR_SchultheisEtAl.pdf}
 }
 

Smart Manufacturing integrates methods of Artificial Intelligence and the Internet of Things into processes to enhance efficiency and flexibility. However, analysis of time series sensor data, crucial for process optimization, is susceptible to Data Quality Issues (DQIs) and can lead to operational problems. Traditional machine learning approaches struggle with limited error data availability in addressing DQIs. The knowledge-driven approach of Case-Based Reasoning targets this issue by reusing experiences regarding already identified DQIs. While some DQIs can be detected using conventional similarity measures, the common, frequently occurring DQI type of Missing Sensor Values pose challenges that cannot be solved using established measures. To address this, this paper proposes a weight-based extension of similarity measures for time series data. This extension aims at the identification and handling of missing sensor values in smart manufacturing processes. Furthermore, analog extensions of established time series measures are presented and possible areas of application outside the DQI domain are outlined.

@inproceedings{Schultheis.2024,
  author       = {Schultheis, Alexander},
  title        = {{Exploring a Hybrid Case-Based Reasoning Approach for Time Series Adaptation in Predictive Maintenance}},
  booktitle    = {Proceedings of the Workshops at the 32nd International Conference
                  on Case-Based Reasoning {(ICCBR-WS} 2024) co-located with the 32nd 
				  International Conference on Case-Based Reasoning {(ICCBR} 2024), Merida,
                  Mexico, July 1, 2024},
  series       = {{CEUR} Workshop Proceedings},
  publisher    = {CEUR-WS.org.},
  editor       = {Lukas Malburg},
  pages        = {230--235},
  volume       = {3708},
  year         = {2024},
  abstract 	   = {{Predictive Maintenance (PredM) is a vital concept within Industry 4.0, focusing on proactive machine maintenance through analysis of sensor data to uphold quality standards and prevent downtime. PredM traditionally employs data analysis methods or Machine Learning (ML) algorithms for anomaly detection in time series data from sensors. Despite ample error-free data, the occurrence of errors is rare. Case-Based Reasoning (CBR) offers an adaptive artificial intelligence approach effective in domains with limited fault data. The sub-research area of Temporal Case-Based Reasoning (TCBR) explores the processing of time series data based on CBR methods. Integrating TCBR methods into PredM leverages human involvement, addressing data privacy concerns and facilitating knowledge transfer. 
  While the retrieval in TCBR has already been investigated, the adaptation of the time series contained in the retrieval results has not yet been considered. On this basis, however, it is possible to determine the further course of the time series as an alternative to ML prediction approaches. For the PredM use case with rare fault data, it is important to determine the further course of the time series and how much time remains before a possible fault case occurs. This research summary therefore investigates a hybrid CBR approach that uses deep learning methods like transformers for adaptation. The aim is to predict the course of a time series as accurately as possible, which is evaluated for the PredM use case. Such a hybrid CBR model should also extend an explanatory component for the predicted time series.}},
  keywords 	   = {{Temporal Case-Based Reasoning, Internet of Things, Time Series Data, Hybrid Case-Based Reasoning, Explainable Case-Based Reasoning, Predictive Maintenance}},
    url = {https://www.wi2.uni-trier.de/shared/publications/2024_ICCBR_DC_Schultheis.pdf}
}

Predictive Maintenance (PredM) is a vital concept within Industry 4.0, focusing on proactive machine maintenance through analysis of sensor data to uphold quality standards and prevent downtime. PredM traditionally employs data analysis methods or Machine Learning (ML) algorithms for anomaly detection in time series data from sensors. Despite ample error-free data, the occurrence of errors is rare. Case-Based Reasoning (CBR) offers an adaptive artificial intelligence approach effective in domains with limited fault data. The sub-research area of Temporal Case-Based Reasoning (TCBR) explores the processing of time series data based on CBR methods. Integrating TCBR methods into PredM leverages human involvement, addressing data privacy concerns and facilitating knowledge transfer. While the retrieval in TCBR has already been investigated, the adaptation of the time series contained in the retrieval results has not yet been considered. On this basis, however, it is possible to determine the further course of the time series as an alternative to ML prediction approaches. For the PredM use case with rare fault data, it is important to determine the further course of the time series and how much time remains before a possible fault case occurs. This research summary therefore investigates a hybrid CBR approach that uses deep learning methods like transformers for adaptation. The aim is to predict the course of a time series as accurately as possible, which is evaluated for the PredM use case. Such a hybrid CBR model should also extend an explanatory component for the predicted time series.

@article{SchultheisEASY2024,
  author       = {Alexander Schultheis and Benjamin Alt and Sebastian Bast and Achim Guldner and David Jilg and Darko Katic and Johannes Mundorf and Tobias Schlagenhauf and Sebastian Weber and Ralph Bergmann and Simon Bergweiler and Lars Creutz and Guido Dartmann and Lukas Malburg and Stefan Naumann and Mahdi Rezapour and Martin Ruskowski},
  title        = {{EASY: Energy-Efficient Analysis and Control Processes in the Dynamic Edge-Cloud Continuum for Industrial Manufacturing}},
  journal      = {K{\"{u}}nstliche Intelligenz},
  year         = {2024},
  keywords 	   = {{Edge-Cloud Continuum, Energy- and Resource-Efficiency, Analysis and Control Processes}},
  abstract 	   = {According to the guiding principles of Industry 4.0, edge computing enables the data-sovereign and near-real-time processing of data directly at the point of origin. Using these edge devices in manufacturing organization will drive the use of industrial analysis, control, and Artificial Intelligence (AI) applications close to production. The goal of the EASY project is to make the added value of edge computing available by providing an easily usable edge-cloud continuum with a runtime environment and services for the execution of AI-based analysis and control processes. Within this continuum, a dynamic, distributed, and optimized execution of services is automated across the entire spectrum from centralized cloud to decentralized edge instances to increase productivity and resource efficiency.},
  doi = {10.1007/s13218-024-00868-3},
    url = {https://www.wi2.uni-trier.de/shared/publications/2024_KI_EASY_SchultheisEtAl.pdf}
}

According to the guiding principles of Industry 4.0, edge computing enables the data-sovereign and near-real-time processing of data directly at the point of origin. Using these edge devices in manufacturing organization will drive the use of industrial analysis, control, and Artificial Intelligence (AI) applications close to production. The goal of the EASY project is to make the added value of edge computing available by providing an easily usable edge-cloud continuum with a runtime environment and services for the execution of AI-based analysis and control processes. Within this continuum, a dynamic, distributed, and optimized execution of services is automated across the entire spectrum from centralized cloud to decentralized edge instances to increase productivity and resource efficiency.

@inproceedings{Seiger.2024_MicroserviceSmartFactorySoftware,
  author       = {Seiger, Ronny and Malburg, Lukas},
  title        = {{Revision of a Smart Factory Software Architecture from Monolith to Microservices}},
  booktitle    = {{Enterprise Design, Operations, and Computing. {EDOC} 2024 Forum, Vienna, Austria, September 10-13, 2024}},
  series       = {Lecture Notes in Business Information Processing},
  %pages       = {50--66},
  %volume      = {14775},
  %doi         = {10.1007/978-3-031-63646-2\_4},
  publisher    = {Springer.},
  note		   = {{Accepted for Publication.}},
  year         = {2024},
  abstract 	   = {{Software architecture plays an important role in the development of modern, complex software systems as it influences a system’s quality attributes and ability to grow with future demand. Designing the software architecture of cyber-physical systems (CPS) becomes even more challenging due to their capability of directly influencing the physical world and thus introducing new non-functional requirements related to fault-tolerance, safety, and resource scarcity. Existing research focuses on systems engineering to achieve the vertical integration of CPS with an organization’s information systems and processes, but not on software architecture to horizontally extend existing systems with new CPS. In this report we describe the process of revising an existing monolithic software architecture for a smart factory towards a microservices-based architecture to meet these new requirements and prepare the factory to be extended with new CPS. For the revision of the existing architecture, we provide an analysis of its code base before and after changes, a description of the refactoring process, and discuss relevant new nonfunctional requirements and architecture options. We elaborate on the architectural decisions favoring microservices and analyze the new architecture regarding improved quality attributes to evaluate the system.}},
  keywords 	   = {{Cyber-physical Systems, Software Architecture, Internet of Things, Microservices, Industry 4.0}},
  url = {https://www.wi2.uni-trier.de/shared/publications/2024_EDOC_MicroserviceSmartFactorySoftware_SeigerMalburg.pdf}
}

Software architecture plays an important role in the development of modern, complex software systems as it influences a system’s quality attributes and ability to grow with future demand. Designing the software architecture of cyber-physical systems (CPS) becomes even more challenging due to their capability of directly influencing the physical world and thus introducing new non-functional requirements related to fault-tolerance, safety, and resource scarcity. Existing research focuses on systems engineering to achieve the vertical integration of CPS with an organization’s information systems and processes, but not on software architecture to horizontally extend existing systems with new CPS. In this report we describe the process of revising an existing monolithic software architecture for a smart factory towards a microservices-based architecture to meet these new requirements and prepare the factory to be extended with new CPS. For the revision of the existing architecture, we provide an analysis of its code base before and after changes, a description of the refactoring process, and discuss relevant new nonfunctional requirements and architecture options. We elaborate on the architectural decisions favoring microservices and analyze the new architecture regarding improved quality attributes to evaluate the system.

@article{Schultheis.2024_Processes,
  author       = {Schultheis, Alexander and Jilg, David and Malburg, Lukas and Bergweiler, Simon and Bergmann, Ralph},
  title        = {{Towards Flexible Control of Production Processes: A Requirements Analysis for Adaptive Workflow Management and Evaluation of Suitable Process Modeling Languages}},
  journal      = {Processes},
  volume 	   = {12},
  year 		   = {2024},
  number 	   = {12},
  article-number = {2714},
  year         = {2024},
  keywords 	   = {{Industry 4.0; Business Process Management; Adaptive Workflow Management; Imperative Process Modeling Languages; Flexible Control Processes}},
  abstract 	   = {In the context of Industry 4.0, Artificial Intelligence (AI) methods are used to maximize the efficiency and flexibility of production processes. The adaptive management of such semantic processes can optimize energy and resource efficiency while providing high reliability, but it depends on the representation type of these models. This paper provides a literature review of current Process Modeling Languages (PMLs). Based on a suitable PML, the flexibility of production processes can be increased. Currently, a common understanding of this process flexibility in the context of adaptive workflow management is missing. Therefore, requirements derived from the business environment are presented for process flexibility. To enable the identification of suitable PLMs, requirements regarding this are also raised. Based on these, the PMLs identified in the literature review are evaluated. Thereby, based on a preselection, a detailed examination of the seven most promising languages is performed, including an example from a real smart factory. As a result, a recommendation is made for the use of BPMN, for which it is presented how it can be enriched with separate semantic information that is suitable for the use of AI planning and, thus, enables flexible control.},
  doi = {10.3390/pr12122714},
  url = {https://www.wi2.uni-trier.de/shared/publications/2024_Processes_SchultheisEtAl.pdf}
}

In the context of Industry 4.0, Artificial Intelligence (AI) methods are used to maximize the efficiency and flexibility of production processes. The adaptive management of such semantic processes can optimize energy and resource efficiency while providing high reliability, but it depends on the representation type of these models. This paper provides a literature review of current Process Modeling Languages (PMLs). Based on a suitable PML, the flexibility of production processes can be increased. Currently, a common understanding of this process flexibility in the context of adaptive workflow management is missing. Therefore, requirements derived from the business environment are presented for process flexibility. To enable the identification of suitable PLMs, requirements regarding this are also raised. Based on these, the PMLs identified in the literature review are evaluated. Thereby, based on a preselection, a detailed examination of the seven most promising languages is performed, including an example from a real smart factory. As a result, a recommendation is made for the use of BPMN, for which it is presented how it can be enriched with separate semantic information that is suitable for the use of AI planning and, thus, enables flexible control.

@article{Malburg_MAPEK_Loops_2023,
	title        = {{Applying MAPE-K control loops for adaptive workflow management in smart factories}},
	author       = {Lukas Malburg and Maximilian Hoffmann and Ralph Bergmann},
	year         = 2023,
	journal      = {{Journal of Intelligent Information Systems}},
	pages		 = {83--111},
	volume       = {61},
	number       = {1},
	doi          = {10.1007/s10844-022-00766-w},
	url          = {http://www.wi2.uni-trier.de/shared/publications/2023_MalburgEtAl_MAPEK_Loops.pdf},
	keywords     = {{Complex event processing, Automated planning, Cyber-physical environments, Smart factories, Adaptive workflow management, Process adaptation}},
	abstract     = {Monitoring the state of currently running processes and reacting to ad-hoc situations during runtime is a key challenge in Business Process Management (BPM). This is especially the case in cyber-physical environments that are characterized by high context sensitivity. MAPE-K control loops are widely used for self-management in these environments and describe four phases for approaching this challenge: Monitor, Analyze, Plan, and Execute. In this paper, we present an architectural solution as well as implementation proposals for using MAPE-K control loops for adaptive workflow management in smart factories. We use Complex Event Processing (CEP) techniques and the process execution states of a Workflow Management System (WfMS) in the monitoring phase. In addition, we apply automated planning techniques to resolve detected exceptional situations and to continue process execution. The experimental evaluation with a physical smart factory shows the potential of the developed approach that is able to detect failures by using IoT sensor data and to resolve them autonomously in near real time with considerable results.}
}

Monitoring the state of currently running processes and reacting to ad-hoc situations during runtime is a key challenge in Business Process Management (BPM). This is especially the case in cyber-physical environments that are characterized by high context sensitivity. MAPE-K control loops are widely used for self-management in these environments and describe four phases for approaching this challenge: Monitor, Analyze, Plan, and Execute. In this paper, we present an architectural solution as well as implementation proposals for using MAPE-K control loops for adaptive workflow management in smart factories. We use Complex Event Processing (CEP) techniques and the process execution states of a Workflow Management System (WfMS) in the monitoring phase. In addition, we apply automated planning techniques to resolve detected exceptional situations and to continue process execution. The experimental evaluation with a physical smart factory shows the potential of the developed approach that is able to detect failures by using IoT sensor data and to resolve them autonomously in near real time with considerable results.

@inproceedings{Malburg.2023_AdaptiveWorkflowsByCBRAndPlanning,
  author    = {Lukas Malburg and
			   Florian Brand and 
               Ralph Bergmann},
  editor    = {Sales, Tiago Prince and Proper, Henderik A. and Guizzardi, Giancarlo and Montali, Marco and Maggi, Fabrizio Maria and Fonseca, Claudenir M.},
  title     = {{Adaptive Management of Cyber-Physical Workflows by Means of Case-Based Reasoning and Automated Planning}},
  booktitle = {Enterprise Design, Operations, and Computing (EDOC) Workshops 2022},
  series    = {Lecture Notes in Business Information Processing},
  volume    = {466},
  pages     = {79--95},
  publisher = {Springer.},
  keywords = {{Case-Based Reasoning, Automated Planning, Industry 4.0, Adaptive Workflow Management, Cyber-Physical Workflows}},
  year      = {2023},
  doi          = {10.1007/978-3-031-26886-1\_5},
  url          = {https://doi.org/10.1007/978-3-031-26886-1\_5},
  abstract = {Today, it is difficult for companies to react to unforeseen events, e.g., global crises. Highly standardized manufacturing processes are particularly limited in their ability to react flexibly, creating a demand for more advanced workflow management techniques, e.g., extended by artificial intelligence methods. In this paper, we describe how Case-Based Reasoning (CBR) can be combined with automated planning to enhance flexibility in cyber-physical production workflows. We present a compositional adaptation method complemented with generative adaptation to resolve unexpected situations during workflow execution. This synergy is advantageous since CBR provides specific knowledge about already experienced situations, whereas planning assists with general knowledge about the domain. In an experimental evaluation, we show that CBR offers a good basis by reusing cases and by adapting them to better suit the current problem. The combination with automated planning further improves these results and, thus, contributes to enhance the flexibility of cyber-physical workflows.},
  url = {http://www.wi2.uni-trier.de/shared/publications/2023_EDOC_MalburgEtAl_AdaptiveWorkflows_by_CBR_and_Planning.pdf},
  note = {The original publication is available at www.springerlink.com}
}

Today, it is difficult for companies to react to unforeseen events, e.g., global crises. Highly standardized manufacturing processes are particularly limited in their ability to react flexibly, creating a demand for more advanced workflow management techniques, e.g., extended by artificial intelligence methods. In this paper, we describe how Case-Based Reasoning (CBR) can be combined with automated planning to enhance flexibility in cyber-physical production workflows. We present a compositional adaptation method complemented with generative adaptation to resolve unexpected situations during workflow execution. This synergy is advantageous since CBR provides specific knowledge about already experienced situations, whereas planning assists with general knowledge about the domain. In an experimental evaluation, we show that CBR offers a good basis by reusing cases and by adapting them to better suit the current problem. The combination with automated planning further improves these results and, thus, contributes to enhance the flexibility of cyber-physical workflows.

@article{Mangler_DataStreamExtension_2023,
author = {Mangler, Juergen and Grüger, Joscha and Malburg, Lukas and Ehrendorfer, Matthias and Bertrand, Yannis and Benzin, Janik-Vasily and Rinderle-Ma, Stefanie and Serral Asensio, Estefania and Bergmann, Ralph},
title = {{DataStream XES Extension: Embedding IoT Sensor Data into Extensible Event Stream Logs}},
journal = {Future Internet},
volume = {15},
year = {2023},
number = {3},
abstract = {{The Internet of Things (IoT) has been shown to be very valuable for Business Process Management (BPM), for example, to better track and control process executions. While IoT actuators can automatically trigger actions, IoT sensors can monitor the changes in the environment and the humans involved in the processes. These sensors produce large amounts of discrete and continuous data streams, which hold the key to understanding the quality of the executed processes. However, to enable this understanding, it is needed to have a joint representation of the data generated by the process engine executing the process, and the data generated by the IoT sensors. In this paper, we present an extension of the event log standard format XES called DataStream. DataStream enables the connection of IoT data to process events, preserving the full context required for data analysis, even when scenarios or hardware artifacts are rapidly changing. The DataStream extension is designed based on a set of goals and evaluated by creating two datasets for real-world scenarios from the transportation/logistics and manufacturing domains.}},
doi = {10.3390/fi15030109},
  keywords = {{Process Management, Industry 4.0, IoT data, Process Mining, XES}},
  url = {http://www.wi2.uni-trier.de/shared/publications/2023_ManglerEtAl_DataStreamExtension.pdf}
}

The Internet of Things (IoT) has been shown to be very valuable for Business Process Management (BPM), for example, to better track and control process executions. While IoT actuators can automatically trigger actions, IoT sensors can monitor the changes in the environment and the humans involved in the processes. These sensors produce large amounts of discrete and continuous data streams, which hold the key to understanding the quality of the executed processes. However, to enable this understanding, it is needed to have a joint representation of the data generated by the process engine executing the process, and the data generated by the IoT sensors. In this paper, we present an extension of the event log standard format XES called DataStream. DataStream enables the connection of IoT data to process events, preserving the full context required for data analysis, even when scenarios or hardware artifacts are rapidly changing. The DataStream extension is designed based on a set of goals and evaluated by creating two datasets for real-world scenarios from the transportation/logistics and manufacturing domains.

@article{Malburg_ConvertingSWSToPDDL_2023,
	title = {Converting semantic web services into formal planning domain descriptions to enable manufacturing process planning and scheduling in industry 4.0},
	journal = {Engineering Applications of Artificial Intelligence},
	volume = {126},
	pages = {106727},
	year = {2023},
	issn = {0952-1976},
	doi = {10.1016/j.engappai.2023.106727},
	author = {Lukas Malburg and Patrick Klein and Ralph Bergmann},
	keywords = {Semantic web services, Industry 4.0, Automated planning, Planning domain definition language, Cyber-physical workflows},
	abstract = {To build intelligent manufacturing systems that react flexibly in case of failures or unexpected circumstances, manufacturing capabilities of production systems must be utilized as much as possible. Artificial Intelligence (AI) and, in particular, automated planning can contribute to this by enabling flexible production processes. To efficiently leverage automated planning, an almost complete planning domain description of the real-world is necessary. However, creating such planning descriptions is a demanding and error-prone task that requires high manual efforts even for domain experts. In addition, maintaining the encoded knowledge is laborious and, thus, can lead to outdated domain descriptions. To reduce the high efforts, already existing knowledge can be reused and transformed automatically into planning descriptions to benefit from organization-wide knowledge engineering activities. This paper presents a novel approach that reduces the described efforts by reusing existing knowledge for planning and scheduling in Industry 4.0 (I4.0). For this purpose, requirements for developing a converter that transforms existing knowledge are derived from literature. Based on these requirements, the SWS2PDDL converter is developed that transforms the knowledge into formal Planning Domain Definition Language (PDDL) descriptions. The approach’s usefulness is verified by a practical evaluation with a near real-world application scenario by generating failures in a physical smart factory and evaluating the generated re-planned production processes. When comparing the resulting plan quality to those achieved by using a manually modeled planning domain by a domain expert, the automatic transformation by SWS2PDDL leads to comparable or even better results without requiring the otherwise high manual modeling efforts.},
	url = {http://www.wi2.uni-trier.de/shared/publications/2023_Malburg_ConvertingSemanticServicesToPDDL.pdf}
}

To build intelligent manufacturing systems that react flexibly in case of failures or unexpected circumstances, manufacturing capabilities of production systems must be utilized as much as possible. Artificial Intelligence (AI) and, in particular, automated planning can contribute to this by enabling flexible production processes. To efficiently leverage automated planning, an almost complete planning domain description of the real-world is necessary. However, creating such planning descriptions is a demanding and error-prone task that requires high manual efforts even for domain experts. In addition, maintaining the encoded knowledge is laborious and, thus, can lead to outdated domain descriptions. To reduce the high efforts, already existing knowledge can be reused and transformed automatically into planning descriptions to benefit from organization-wide knowledge engineering activities. This paper presents a novel approach that reduces the described efforts by reusing existing knowledge for planning and scheduling in Industry 4.0 (I4.0). For this purpose, requirements for developing a converter that transforms existing knowledge are derived from literature. Based on these requirements, the SWS2PDDL converter is developed that transforms the knowledge into formal Planning Domain Definition Language (PDDL) descriptions. The approach’s usefulness is verified by a practical evaluation with a near real-world application scenario by generating failures in a physical smart factory and evaluating the generated re-planned production processes. When comparing the resulting plan quality to those achieved by using a manually modeled planning domain by a domain expert, the automatic transformation by SWS2PDDL leads to comparable or even better results without requiring the otherwise high manual modeling efforts.

@article{Guldner_AICPPS_2023,
	url = {https://doi.org/10.1515/itit-2023-0001},
	title = {A framework for AI-based self-adaptive cyber-physical process systems},
	author = {Achim Guldner and Maximilian Hoffmann and Christian Lohr and Rüdiger Machhamer and Lukas Malburg and Marlies Morgen and Stephanie C. Rodermund and Florian Schäfer and Lars Schaupeter and Jens Schneider and Felix Theusch and Ralph Bergmann and Guido Dartmann and Norbert Kuhn and Stefan Naumann and Ingo J. Timm and Matthias Vette-Steinkamp and Benjamin Weyers},
	journal = {it - Information Technology},
	volume = {65},
	number = {3},
	url          = {http://www.wi2.uni-trier.de/shared/publications/2023_Guldner_AICPPS.pdf},
	doi = {doi:10.1515/itit-2023-0001},
	keywords     = {{Artificial Intelligence, Business Process Managemenet, Cyber-Physical Systems, Framework, Green AI, Process-Aware Information System}},
	abstract     = {{Digital transformation is both an opportunity and a challenge. To take advantage of this opportunity for humans and the environment, the transformation process must be understood as a design process that affects almost all areas of life. In this paper, we investigate AI-Based Self-Adaptive Cyber-Physical Process Systems (AI-CPPS) as an extension of the traditional CPS view. As contribution, we present a framework that addresses challenges that arise from recent literature. The aim of the AI-CPPS framework is to enable an adaptive integration of IoT environments with higher-level process-oriented systems. In addition, the
framework integrates humans as actors into the system, which is often neglected by recent related approaches. The framework consists of three layers, i.e., processes, semantic modeling, and systems and actors, and we describe for each layer challenges and solution outlines for application. We also address the requirement to enable the integration of new networked devices under the premise of a targeted process that is optimally designed for humans, while profitably integrating AI and IoT. It is expected that AI-CPPS can contribute significantly to increasing sustainability and quality of life and offer solutions to pressing problems such as environmental protection, mobility, or demographic change. Thus, it is all the more important that the systems themselves do not become a driver of resource consumption.}},
	year = {2023}
}

Digital transformation is both an opportunity and a challenge. To take advantage of this opportunity for humans and the environment, the transformation process must be understood as a design process that affects almost all areas of life. In this paper, we investigate AI-Based Self-Adaptive Cyber-Physical Process Systems (AI-CPPS) as an extension of the traditional CPS view. As contribution, we present a framework that addresses challenges that arise from recent literature. The aim of the AI-CPPS framework is to enable an adaptive integration of IoT environments with higher-level process-oriented systems. In addition, the framework integrates humans as actors into the system, which is often neglected by recent related approaches. The framework consists of three layers, i.e., processes, semantic modeling, and systems and actors, and we describe for each layer challenges and solution outlines for application. We also address the requirement to enable the integration of new networked devices under the premise of a targeted process that is optimally designed for humans, while profitably integrating AI and IoT. It is expected that AI-CPPS can contribute significantly to increasing sustainability and quality of life and offer solutions to pressing problems such as environmental protection, mobility, or demographic change. Thus, it is all the more important that the systems themselves do not become a driver of resource consumption.

@article{Grger2023,
  doi = {10.1515/itit-2022-0077},
  url = {https://doi.org/10.1515/itit-2022-0077},
  year = {2023},
  volume = {65},
  number = {3},
  month = jun,
  publisher = {Walter de Gruyter {GmbH}},
  author = {Joscha Gr\"{u}ger and Lukas Malburg and Ralph Bergmann},
  title = {{IoT}-enriched event log generation and quality analytics: a case study},
  url          = {http://www.wi2.uni-trier.de/shared/publications/10.1515_itit-2022-0077.pdf},
  journal = {it - Information Technology},
  keywords     = {{Data Quality in Event Logs, DataStream XES Extension, IoT, IoT-Enriched Event Log, Physical Smart Factory, Process Mining}},
  abstract     = {Modern technologies such as the Internet of Things (IoT) are becoming increasingly important in various fields, including business process management (BPM) research. An important area of research in BPM is process mining, which can be used to analyze event logs e.g., to check the conformance of running processes. However, the data ingested in IoT environments often contain data quality issues (DQIs) due to system complexity and sensor heterogeneity, among other factors. To date, however, there has been little work on IoT event logs, DQIs occurring in them, and how to handle them. In this case study, we generate an IoT event log, perform a structured data quality analysis, and describe how we addressed the problems we encountered in pre-processing.}
} 

Modern technologies such as the Internet of Things (IoT) are becoming increasingly important in various fields, including business process management (BPM) research. An important area of research in BPM is process mining, which can be used to analyze event logs e.g., to check the conformance of running processes. However, the data ingested in IoT environments often contain data quality issues (DQIs) due to system complexity and sensor heterogeneity, among other factors. To date, however, there has been little work on IoT event logs, DQIs occurring in them, and how to handle them. In this case study, we generate an IoT event log, perform a structured data quality analysis, and describe how we addressed the problems we encountered in pre-processing.

@inproceedings{Brand.2023_RLForAdaptiveWorkflows,
  author       = {Florian Brand and
                  Katharina Lott and
                  Lukas Malburg and
                  Maximilian Hoffmann and
                  Ralph Bergmann},
  editor       = {Lukas Malburg and
                  Deepika Verma},
  title        = {{Using Deep Reinforcement Learning for the Adaptation of Semantic Workflows}},
  booktitle    = {Proceedings of the Workshops at the 31st International Conference
                  on Case-Based Reasoning {(ICCBR-WS} 2023) co-located with the 31st
                  International Conference on Case-Based Reasoning {(ICCBR} 2023), Aberdeen,
                  Scotland, UK, July 17, 2023},
  series       = {{CEUR} Workshop Proceedings},
  volume       = {3438},
  pages        = {55--70},
  publisher    = {CEUR-WS.org},
  year         = {2023},
  keywords     = {{Case-Based Reasoning, Semantic Workflows, Deep Learning, Reinforcement Learning}},
  abstract     = {Case-Based Reasoning (CBR) solves new problems by using experience represented by solved cases. The acquisition of adaptation knowledge and its subsequent application remains a classic challenge for CBR applications. In this paper, we present a novel approach for adapting semantic workflows during the reuse phase of the CBR cycle. A reinforcement learning agent is utilized, which applies different actions to change nodes of the workflow. Thereby, changes to the workflow are made by replacing, deleting or adding nodes. The agent is evaluated in a case study outlining its ability to adapt a semantic graph in a smart manufacturing domain. While the approach is detailed for the application in the particular domain, it can be adopted for the usage in other process-oriented domains.},
  url          = {http://www.wi2.uni-trier.de/shared/publications/2023_Brand_RLForAdaptiveWorkflows.pdf}
}

Case-Based Reasoning (CBR) solves new problems by using experience represented by solved cases. The acquisition of adaptation knowledge and its subsequent application remains a classic challenge for CBR applications. In this paper, we present a novel approach for adapting semantic workflows during the reuse phase of the CBR cycle. A reinforcement learning agent is utilized, which applies different actions to change nodes of the workflow. Thereby, changes to the workflow are made by replacing, deleting or adding nodes. The agent is evaluated in a case study outlining its ability to adapt a semantic graph in a smart manufacturing domain. While the approach is detailed for the application in the particular domain, it can be adopted for the usage in other process-oriented domains.

@inproceedings{Malburg.2023_TimeSeriesInCBR,
  author       = {Lukas Malburg and
                  Alexander Schultheis and
                  Ralph Bergmann},
  editor       = {Lukas Malburg and
                  Deepika Verma},
  title        = {{Modeling and Using Complex IoT Time Series Data in Case-Based Reasoning:
                  From Application Scenarios to Implementations}},
  booktitle    = {Proceedings of the Workshops at the 31st International Conference
                  on Case-Based Reasoning {(ICCBR-WS} 2023) co-located with the 31st
                  International Conference on Case-Based Reasoning {(ICCBR} 2023), Aberdeen,
                  Scotland, UK, July 17, 2023},
  series       = {{CEUR} Workshop Proceedings},
  volume       = {3438},
  pages        = {81--96},
  publisher    = {CEUR-WS.org},
  year         = {2023},
  keywords     = {{Case-Based Reasoning, Temporal Case-Based Reasoning, Internet of Things, Time Series Data, ProCAKE}},
  abstract     = {The research area of Internet of Things (IoT) is gaining more relevance for several domains and application areas, including Case-Based Reasoning (CBR). However, IoT data is characterized by high volumes and variance of data types, making the application of CBR methods difficult. Since only few works have been published in this area so far, the integration and consideration of complex IoT data such as time series data in CBR frameworks is still in its infancy. To catch up with the current state-of-the-art, we present a comprehensive literature review on Temporal Case-Based Reasoning and time series data in CBR as part of our contribution. Furthermore, we present typical application scenarios for using IoT time series data in practice that can be addressed in further research. To build suitable CBR implementations for that purpose, we define a procedure model that can be used for time series data in CBR. In this context, we address the implementation of the application scenarios in the ProCAKE CBR framework.},
  url          = {http://www.wi2.uni-trier.de/shared/publications/2023_Malburg_TimeSeriesInCBR.pdf}
}

The research area of Internet of Things (IoT) is gaining more relevance for several domains and application areas, including Case-Based Reasoning (CBR). However, IoT data is characterized by high volumes and variance of data types, making the application of CBR methods difficult. Since only few works have been published in this area so far, the integration and consideration of complex IoT data such as time series data in CBR frameworks is still in its infancy. To catch up with the current state-of-the-art, we present a comprehensive literature review on Temporal Case-Based Reasoning and time series data in CBR as part of our contribution. Furthermore, we present typical application scenarios for using IoT time series data in practice that can be addressed in further research. To build suitable CBR implementations for that purpose, we define a procedure model that can be used for time series data in CBR. In this context, we address the implementation of the application scenarios in the ProCAKE CBR framework.

@inproceedings{Ehrendorfer.2023_IoPTRepository,
  author       = {Matthias Ehrendorfer and
                  Yannis Bertrand and
                  Lukas Malburg and
                  Juergen Mangler and
                  Joscha Gr{\"{u}}ger and
                  Stefanie Rinderle{-}Ma and
                  Ralph Bergmann and
                  Estefan{\'{\i}}a Serral Asensio},
  editor       = {Dirk Fahland and
                  Andr{\'{e}}s Jim{\'{e}}nez{-}Ram{\'{\i}}rez and
                  Akhil Kumar and
                  Jan Mendling and
                  Brian T. Pentland and
                  Stefanie Rinderle{-}Ma and
                  Tijs Slaats and
                  Johan Versendaal and
                  Barbara Weber and
                  Mathias Weske and
                  Karolin Winter},
  title        = {{Internet of Processes and Things: {A} Repository for IoT-Enriched
                  Event Logs in Smart Environments}},
  booktitle    = {{Proceedings of the Best Dissertation Award, Doctoral Consortium, and Demonstration {\&} Resources Forum at {BPM} 2023 co-located with 21st International Conference on Business Process Management {(BPM} 2023), Utrecht, The Netherlands, September 11th to 15th, 2023}},
  series       = {{CEUR} Workshop Proceedings},
  volume       = {3469},
  pages        = {92--96},
  keywords     = {{IoT-Enriched Event Logs, Business Process Management, IoT}},
  abstract     = {Current research efforts from the Business Processes Management and IoT communities, touching novel techniques such as object-centric process mining, predictive process monitoring or IoT based process enhancement suffer from a lack of real-world open data-sets from different domains to evaluate and refine their approaches. In this paper we present an online resource for the collection of IoT-enhanced process logs, hoping to inspire companies, organisations and researchers to cooperate in the release of well-understood high-quality data sets.},
  publisher    = {CEUR-WS.org},
  year         = {2023},
  url          = {http://www.wi2.uni-trier.de/shared/publications/2023_Ehrendorfer_IoPTRepository.pdf}
}

Current research efforts from the Business Processes Management and IoT communities, touching novel techniques such as object-centric process mining, predictive process monitoring or IoT based process enhancement suffer from a lack of real-world open data-sets from different domains to evaluate and refine their approaches. In this paper we present an online resource for the collection of IoT-enhanced process logs, hoping to inspire companies, organisations and researchers to cooperate in the release of well-understood high-quality data sets.

@inproceedings{Hoffmann.2022_ProGAN,
  author    = {Maximilian Hoffmann and
			   Lukas Malburg and
               Ralph Bergmann},
  editor    = {Andrea Marrella and Barbara Weber},
  title     = {{ProGAN: Toward a Framework for Process Monitoring and Flexibility by Change via Generative Adversarial Networks}},
  booktitle = {Business Process Management Workshops - {BPM} 2021,
               Rome, Italy, September 6 - 10, 2021},
  series    = {Lecture Notes in Business Information Processing},
  volume    = {436},
  pages     = {43--55},
  publisher = {Springer.},
  keywords = {{Business process prediction, Generative Adversarial Networks, Flexibility by change, Process adaptation}},
  year      = {2022},
  url       = {https://doi.org/10.1007/978-3-030-94343-1\_4},
  doi       = {10.1007/978-3-030-94343-1\_4},
  abstract = {Monitoring the state of currently running processes and reacting to deviations during runtime is a key challenge in Business Process Management (BPM). The MAPE-K control loop describes four phases for approaching this challenge: Monitor, Analyze, Plan, Execute. In this paper, we present the ProGAN framework, an idea of an approach for implementing the monitor, analyze, and plan phases of MAPE-K. For this purpose, we leverage a deep learning architecture that builds upon Generative Adversarial Networks (GANs): The discriminator is used for monitoring the process in its environment by using sensor data and for detecting deviations w.r.t. the desired process state (monitor phase). The generator is used afterwards for analyzing the detected deviation and its symptoms as well as for adapting the current process to resolve the deviation and to restore the desired state. Both components are trained together by utilizing each other's feedback in a self-supervised way. We demonstrate the application of our approach for an exemplary scenario in the manufacturing domain.},
  url = {http://www.wi2.uni-trier.de/shared/publications/2022_AI4BPM_HoffmannEtal_ProGAN.pdf},
  note = {The original publication is available at www.springerlink.com}
}

Monitoring the state of currently running processes and reacting to deviations during runtime is a key challenge in Business Process Management (BPM). The MAPE-K control loop describes four phases for approaching this challenge: Monitor, Analyze, Plan, Execute. In this paper, we present the ProGAN framework, an idea of an approach for implementing the monitor, analyze, and plan phases of MAPE-K. For this purpose, we leverage a deep learning architecture that builds upon Generative Adversarial Networks (GANs): The discriminator is used for monitoring the process in its environment by using sensor data and for detecting deviations w.r.t. the desired process state (monitor phase). The generator is used afterwards for analyzing the detected deviation and its symptoms as well as for adapting the current process to resolve the deviation and to restore the desired state. Both components are trained together by utilizing each other's feedback in a self-supervised way. We demonstrate the application of our approach for an exemplary scenario in the manufacturing domain.

@article{Grueger.2022_SensorStreamExtension,
  author    = {Joscha Grüger and Lukas Malburg and Jürgen Mangler and Yannis Bertrand and Stefanie Rinderle-Ma and Ralph Bergmann and Estefanía {Serral Asensio}},
  title     = {{SensorStream: An XES Extension for Enriching Event Logs with IoT-Sensor Data}},
  journal   = {CoRR},
  volume    = {abs/2206.11392},
  year      = {2022},
  url       = {https://arxiv.org/abs/2206.11392},
  archivePrefix = {arXiv},
  eprint    = {2206.11392},
  abstract = {Process management and process orchestration/execution are currently hot topics; prevalent trends such as automation and Industry 4.0 require solutions which allow domain-experts to easily model and execute processes in various domains, including manufacturing and health-care. These domains, in turn, rely on a tight integration between hardware and software, i.e. via the Internet of Things (IoT). While process execution is about actuation, i.e. actively triggering actions and awaiting their completion, accompanying IoT sensors monitor humans and the environment. These sensors produce large amounts of procedural, discrete, and continuous data streams, that hold the key to understanding the quality of process subjects (e.g. produced parts), outcome (e.g. quantity and quality), and error causes. Processes constantly evolve in conjunction with their IoT environment. This requires joint storage of data generated by processes, with data generated by the IoT sensors is therefore needed. In this paper, we present an extension of the process log standard format XES, namely SensorStream. SensorStream enables to connect IoT data to process events, as well as a set of semantic annotations to describe the scenario and environment during data collection. This allows to preserve the full context required for data-analysis, so that logs can be analyzed even when scenarios or hardware artifacts are rapidly changing. Through additional semantic annotations, we envision the XES extension log format to be a solid based for the creation of a (semi-)automatic analysis pipeline, which can support domain experts by automatically providing data visualization, or even process insights.}
}

Process management and process orchestration/execution are currently hot topics; prevalent trends such as automation and Industry 4.0 require solutions which allow domain-experts to easily model and execute processes in various domains, including manufacturing and health-care. These domains, in turn, rely on a tight integration between hardware and software, i.e. via the Internet of Things (IoT). While process execution is about actuation, i.e. actively triggering actions and awaiting their completion, accompanying IoT sensors monitor humans and the environment. These sensors produce large amounts of procedural, discrete, and continuous data streams, that hold the key to understanding the quality of process subjects (e.g. produced parts), outcome (e.g. quantity and quality), and error causes. Processes constantly evolve in conjunction with their IoT environment. This requires joint storage of data generated by processes, with data generated by the IoT sensors is therefore needed. In this paper, we present an extension of the process log standard format XES, namely SensorStream. SensorStream enables to connect IoT data to process events, as well as a set of semantic annotations to describe the scenario and environment during data collection. This allows to preserve the full context required for data-analysis, so that logs can be analyzed even when scenarios or hardware artifacts are rapidly changing. Through additional semantic annotations, we envision the XES extension log format to be a solid based for the creation of a (semi-)automatic analysis pipeline, which can support domain experts by automatically providing data visualization, or even process insights.

@article{Malburg.2022_IoTEnrichedEventLog,
  author    = {Lukas Malburg and Joscha Grüger and Ralph Bergmann},
  title     = {{An IoT-Enriched Event Log for Process Mining in Smart Factories}},
  journal   = {CoRR},
  volume    = {abs/2209.02702},
  year      = {2022},
  url       = {https://arxiv.org/abs/2209.02702},
  archivePrefix = {arXiv},
  eprint    = {2209.02702},
  abstract = {Modern technologies such as the Internet of Things (IoT) are becoming increasingly important in various domains, including Business Process Management (BPM) research. One main research area in BPM is process mining, which can be used to analyze event logs, e.g., for checking the conformance of running processes. However, there are only a few IoT-based event logs available for research purposes. Some of them are artificially generated and the problem occurs that they do not always completely reflect the actual physical properties of smart environments. In this paper, we present an IoT-enriched XES event log that is generated by a physical smart factory. For this purpose, we create the SensorStream XES extension for representing IoT-data in event logs. Finally, we present some preliminary analysis and properties of the log.}
}

Modern technologies such as the Internet of Things (IoT) are becoming increasingly important in various domains, including Business Process Management (BPM) research. One main research area in BPM is process mining, which can be used to analyze event logs, e.g., for checking the conformance of running processes. However, there are only a few IoT-based event logs available for research purposes. Some of them are artificially generated and the problem occurs that they do not always completely reflect the actual physical properties of smart environments. In this paper, we present an IoT-enriched XES event log that is generated by a physical smart factory. For this purpose, we create the SensorStream XES extension for representing IoT-data in event logs. Finally, we present some preliminary analysis and properties of the log.

@article{seiger_process_management_smart_factories_2022,
 author = {Seiger, Ronny and Malburg, Lukas and Weber, Barbara and Bergmann, Ralph},
 year = {2022},
 title = {{Integrating process management and event processing in smart factories: A systems architecture and use cases}},
 pages = {575--592},
 volume = {63},
 abstract = {The developments of new concepts for an increased digitization of manufacturing industries in the context of Industry 4.0 have brought about novel system architectures and frameworks for smart production systems. These range from generic frameworks for Industry 4.0 to domain-specific architectures for Industrial Internet of Things (IIoT). While most of the approaches include a service-based architecture for selective integration with enterprise systems, a close two-way integration of the production control systems and IIoT sensors and actuators with Process-Aware Information Systems (PAIS) on the management level for automation and mining of production processes is rarely discussed. This fusion of Business Process Management (BPM) with IIoT can be mutually beneficial for both research areas, but is still in its infancy. We propose a systems architecture for IIoT that shows how to integrate the low-level hardware components – sensors and actuators – of a smart factory with BPM systems. We discuss the software components and their interactions to address challenges of device encapsulation, integration of sensor events, and interaction with existing BPM systems. This integration is demonstrated within several use cases regarding process modeling, automation and mining for a smart factory model, showing benefits of using BPM technologies to analyze, control, and adapt discrete production processes in IIoT.},
 journal = {{Journal of Manufacturing Systems}},
 url = {http://www.wi2.uni-trier.de/shared/publications/2022_Seiger_JOMS.pdf},
 doi = {10.1016/j.jmsy.2022.05.012}
}

The developments of new concepts for an increased digitization of manufacturing industries in the context of Industry 4.0 have brought about novel system architectures and frameworks for smart production systems. These range from generic frameworks for Industry 4.0 to domain-specific architectures for Industrial Internet of Things (IIoT). While most of the approaches include a service-based architecture for selective integration with enterprise systems, a close two-way integration of the production control systems and IIoT sensors and actuators with Process-Aware Information Systems (PAIS) on the management level for automation and mining of production processes is rarely discussed. This fusion of Business Process Management (BPM) with IIoT can be mutually beneficial for both research areas, but is still in its infancy. We propose a systems architecture for IIoT that shows how to integrate the low-level hardware components – sensors and actuators – of a smart factory with BPM systems. We discuss the software components and their interactions to address challenges of device encapsulation, integration of sensor events, and interaction with existing BPM systems. This integration is demonstrated within several use cases regarding process modeling, automation and mining for a smart factory model, showing benefits of using BPM technologies to analyze, control, and adapt discrete production processes in IIoT.

@inproceedings{Malburg_AdaptiveWorkflowManagement_2022,
	title        = {{Towards Adaptive Workflow Management by Case-Based Reasoning and Automated Planning}},
	author       = {Malburg, Lukas and Bergmann, Ralph},
	year         = 2022,
	booktitle    = {{Workshops Proceedings for the Thirtieth International Conference on Case-Based Reasoning co-located with the Thirtieth International Conference on Case-Based Reasoning {(ICCBR} 2022), Nancy, France, September 12-15, 2022}},
	publisher    = {CEUR-WS.org},
	volume       = {3389},
	pages        = {211--220},
	series       = {{CEUR} Workshop Proceedings},
	editor       = {Pascal Reuss and Jakob Schönborn},
	url          = {http://www.wi2.uni-trier.de/shared/publications/2023_Malburg_ICCBR.pdf},
	keywords     = {{Case-Based Reasoning, Automated Planning, Industry 4.0, Adaptive Workflow Management, Cyber-Physical Workflows}},
	abstract     = {Adaptive workflow management is an important topic in recent years, as increasing dynamics due to growing customer demands and faster changing market conditions require more flexibility in workflows. This is especially the case for rigid and rather standardized production processes that cannot be easily modified, if, for example, a breakdown occurs in one of the production lines. The goal of the Fourth Industrial Revolution (Industry 4.0), is to provide, among others, more flexible and cost-effective processes in companies by using Artificial Intelligence (AI) methods. In this paper, we present 1) a framework for adaptive workflow management for IoT-enhanced manufacturing processes and 2) an idea for a new adaptation method that combines Case-Based Reasoning (CBR) and automated planning. In this context, we discuss the benefits of such a synergistic combination and introduce the framework and the individual phases according to the 4R (Retrieve, Reuse, Revise, Retain) CBR cycle. In addition, we present our physical factory simulation model that can be used to evaluate the suitability of developed research artifacts.}
}

Adaptive workflow management is an important topic in recent years, as increasing dynamics due to growing customer demands and faster changing market conditions require more flexibility in workflows. This is especially the case for rigid and rather standardized production processes that cannot be easily modified, if, for example, a breakdown occurs in one of the production lines. The goal of the Fourth Industrial Revolution (Industry 4.0), is to provide, among others, more flexible and cost-effective processes in companies by using Artificial Intelligence (AI) methods. In this paper, we present 1) a framework for adaptive workflow management for IoT-enhanced manufacturing processes and 2) an idea for a new adaptation method that combines Case-Based Reasoning (CBR) and automated planning. In this context, we discuss the benefits of such a synergistic combination and introduce the framework and the individual phases according to the 4R (Retrieve, Reuse, Revise, Retain) CBR cycle. In addition, we present our physical factory simulation model that can be used to evaluate the suitability of developed research artifacts.

@INPROCEEDINGS{klein_SiameseNetworkPredM_2021,
  author={Patrick Klein and Niklas Weingarz and Ralph Bergmann},
  booktitle={2021 International Joint Conference on Neural Networks (IJCNN)}, 
  title={Using Expert Knowledge for Masking Irrelevant Data Streams in Siamese Networks for the Detection and Prediction of Faults}, 
  year={2021},
  note={Accepted for presentation.}}

@article{malburg_objectDetection_2021,
  author    = {Lukas Malburg and Manfred-Peter Rieder and Ronny Seiger and Patrick Klein and Ralph Bergmann},
  title     = {{Object Detection for Smart Factory Processes by Machine Learning}},
  booktitle = {The 4th International Conference on Emerging Data and Industry 4.0 (EDI40), Warsaw, Poland, March 23 - 26, 2021},
  journal    = {Procedia Computer Science},
  volume    = {184},
  pages     = {581--588},
  publisher = {Elsevier.},
  keywords = {{Process Monitoring, Object Detection, Computer Vision, Machine Learning, Industry 4.0, Cyber-Physical Production Systems}},
  year      = {2021},
  url       = {https://doi.org/10.1016/j.procs.2021.04.009},
  doi       = {10.1016/j.procs.2021.04.009},
  abstract={The production industry is in a transformation towards more autonomous and intelligent manufacturing. In addition to more flexible production processes to dynamically respond to changes in the environment, it is also essential that production processes are continuously monitored and completed in time. Video-based methods such as object detection systems are still in their infancy and rarely used as basis for process monitoring. In this paper, we present a framework for video-based monitoring of manufacturing processes with the help of a physical smart factory simulation model. We evaluate three state-of-the-art object detection systems regarding their suitability to detect workpieces and to recognize failure situations that require adaptations. In our experiments, we are able to show that detection accuracies above 90% can be achieved with current object detection methods.},
  url = {http://www.wi2.uni-trier.de/shared/publications/2021_MalburgEtAl_ObjectDetectionInSmartFactories.pdf},
  note = {Best Paper.}
}

The production industry is in a transformation towards more autonomous and intelligent manufacturing. In addition to more flexible production processes to dynamically respond to changes in the environment, it is also essential that production processes are continuously monitored and completed in time. Video-based methods such as object detection systems are still in their infancy and rarely used as basis for process monitoring. In this paper, we present a framework for video-based monitoring of manufacturing processes with the help of a physical smart factory simulation model. We evaluate three state-of-the-art object detection systems regarding their suitability to detect workpieces and to recognize failure situations that require adaptations. In our experiments, we are able to show that detection accuracies above 90% can be achieved with current object detection methods.

@inproceedings{malburg_BPMResearch_2020,
  author    = {Lukas Malburg and
               Ronny Seiger and
               Ralph Bergmann and
			   Barbara Weber},
  editor    = {Adela del-Río-Ortega and Henrik Leopold and Flavia M. Santoro},
  title     = {{Using Physical Factory Simulation Models for Business Process Management Research}},
  booktitle = {Business Process Management Workshops - {BPM} 2020 International Workshops,
               Sevilla, Spain, September 13 - 18, 2020},
  series    = {Lecture Notes in Business Information Processing},
  volume    = {397},
  pages     = {95--107},
  publisher = {Springer},
  keywords = {{Cyber-Physical Production Systems, Factory Simulation Models, Business Process Management, Industry 4.0, Digital Twins}},
  year      = {2020},
  url       = {https://doi.org/10.1007/978-3-030-66498-5\_8},
  doi       = {10.1007/978-3-030-66498-5\_8},
  abstract={The production and manufacturing industries are currently transitioning towards more autonomous and intelligent production lines within the Fourth Industrial Revolution (Industry 4.0). Learning Factories as small scale physical models of real shop floors are realistic platforms to conduct research in the smart manufacturing area without depending on expensive real world production lines or completely simulated data. In this work, we propose to use learning factories for conducting research in the context of Business Process Management (BPM) and Internet of Things (IoT) as this combination promises to be mutually beneficial for both research areas. We introduce our physical Fischertechnik factory models simulating a complex production line and three exemplary use cases of combining BPM and IoT, namely the implementation of a BPM abstraction stack on top of a learning factory, the experience-based adaptation and optimization of manufacturing processes, and the stream processing-based conformance checking of IoT-enabled processes.},
  url = {http://www.wi2.uni-trier.de/shared/publications/2020_MalburgEtAl_BPM.pdf},
  note = {The original publication is available at www.springerlink.com}
}

The production and manufacturing industries are currently transitioning towards more autonomous and intelligent production lines within the Fourth Industrial Revolution (Industry 4.0). Learning Factories as small scale physical models of real shop floors are realistic platforms to conduct research in the smart manufacturing area without depending on expensive real world production lines or completely simulated data. In this work, we propose to use learning factories for conducting research in the context of Business Process Management (BPM) and Internet of Things (IoT) as this combination promises to be mutually beneficial for both research areas. We introduce our physical Fischertechnik factory models simulating a complex production line and three exemplary use cases of combining BPM and IoT, namely the implementation of a BPM abstraction stack on top of a learning factory, the experience-based adaptation and optimization of manufacturing processes, and the stream processing-based conformance checking of IoT-enabled processes.

@inProceedings{klein_SiameseNN4PredM_2020,
  author    = {Patrick Klein and
               Niklas Weingarz and
               Ralph Bergmann},
  title     = {{Enhancing Siamese Neural Networks through Expert Knowledge for Predictive Maintenance}},
  booktitle = {{IoT Streams for Data-Driven Predictive Maintenance and IoT, Edge, and Mobile for Embedded Machine Learning}},
  series    = {Communications in Computer and Information Science},
  volume    = {1325},
  %pages    = {1--16},
  publisher = {Springer International Publishing.},
  keywords 	= {{Siamese Neural Network, Predictive Maintenance, Expert Knowledge, 2D Convolution, Time Series Similarity}},
  year      = {2020},
  abstract	={The data provided by cyber-physical production systems (CPPSs) to monitor their condition via data-driven predictive maintenance is often high dimensional and only a few fault and failure (FaF) examples are available. These FaFs can usually be detected in a (small) localized subset of data streams, whereas the use of all data streams induces noise that could negatively affect the training and prediction performance. In addition, a CPPS often consists of multiple similar units that generate comparable data streams and show similar failure modes. However, existing approaches for learning a similarity measure generally do not consider these two aspects. For this reason, we propose two approaches for integrating expert knowledge about class or failure mode dependent attributes into siamese neural networks (SNN). Additionally, we present an attribute-wise encoding of time series based on 2D convolutions. This enables that learned knowledge in the form of filters is shared between similar data streams, which would not be possible with conventional 1D convolutions due to their spatial focus. We evaluate our approaches against state-of-the-art time series similarity measures such as dynamic time warping, NeuralWarp, as well as a feature-based representation approach. Our results show that the integration of expert knowledge is advantageous and combined with the novel SNN architecture it is possible to achieve the best performance compared to the other investigated methods..},
  doi       = {10.1007/978-3-030-66770-2\_6},
  isbn		={978-3-030-66769-6},
  year      = {2020},
  url 		= {http://www.wi2.uni-trier.de/shared/publications/2020_ECML-IoTStreams_SiameseNeuralNetwork-for-PredictiveMaintenance_Preprint.pdf},
  note = {The original publication is available at www.springerlink.com}}
  

The data provided by cyber-physical production systems (CPPSs) to monitor their condition via data-driven predictive maintenance is often high dimensional and only a few fault and failure (FaF) examples are available. These FaFs can usually be detected in a (small) localized subset of data streams, whereas the use of all data streams induces noise that could negatively affect the training and prediction performance. In addition, a CPPS often consists of multiple similar units that generate comparable data streams and show similar failure modes. However, existing approaches for learning a similarity measure generally do not consider these two aspects. For this reason, we propose two approaches for integrating expert knowledge about class or failure mode dependent attributes into siamese neural networks (SNN). Additionally, we present an attribute-wise encoding of time series based on 2D convolutions. This enables that learned knowledge in the form of filters is shared between similar data streams, which would not be possible with conventional 1D convolutions due to their spatial focus. We evaluate our approaches against state-of-the-art time series similarity measures such as dynamic time warping, NeuralWarp, as well as a feature-based representation approach. Our results show that the integration of expert knowledge is advantageous and combined with the novel SNN architecture it is possible to achieve the best performance compared to the other investigated methods..

@inproceedings{malburg_SemanticWebServices_2020,
  author    = {Lukas Malburg and
               Patrick Klein and
               Ralph Bergmann},
  title     = {{Semantic Web Services for AI-Research with Physical Factory Simulation Models in Industry 4.0}},
  editor    = {Herv{\'{e}} Panetto and
               Kurosh Madani and
               Alexander V. Smirnov},
  booktitle = {Proceedings of the International Conference on Innovative Intelligent
               Industrial Production and Logistics, {IN4PL} 2020, Budapest, Hungary,
               November 2-4, 2020},
  pages     = {32--43},
  publisher = {{SCITEPRESS.}},
  keywords = {{Semantic Web Services, Industry 4.0, Artificial Intelligence, Flexible Cyber-Physical Workflows, OWL-S, WSMO}},
  year      = {2020},
  abstract={The transition to Industry 4.0 requires smart manufacturing systems that are easily configurable and provide a high level of flexibility during manufacturing in order to achieve mass customization or to support cloud manufacturing. To realize this, Cyber-Physical Systems (CPSs) combined with Artificial Intelligence (AI) methods find their way into manufacturing shop floors. For using AI methods in the context of Industry 4.0, semantic web services are indispensable to provide a reasonable abstraction of the underlying manufacturing capabilities. In this paper, we present semantic web services for AI-based research with physical factory simulation models in Industry 4.0. Therefore, we developed 68 semantic web services based on Web Ontology Language for Web Services (OWL-S) and Web Service Modeling Ontology (WSMO) and linked them to an already existing domain ontology for intelligent manufacturing control. Suitable for the requirements of CPS environments, our pre- and postconditions are verified in near real-time by invoking other semantic web services in contrast to complex reasoning within the knowledge base. Finally, we evaluate our implementation by executing a cyber-physical workflow composed of semantic web services using a state-of-the-art workflow management system.},
  doi={10.5220/0010135900320043},
  pages={32-43},
  isbn={978-989-758-476-3},
  url = {http://www.wi2.uni-trier.de/shared/publications/2020_IN4PL_Semantic_Web_Services_MalburgEtAl.pdf}
}

The transition to Industry 4.0 requires smart manufacturing systems that are easily configurable and provide a high level of flexibility during manufacturing in order to achieve mass customization or to support cloud manufacturing. To realize this, Cyber-Physical Systems (CPSs) combined with Artificial Intelligence (AI) methods find their way into manufacturing shop floors. For using AI methods in the context of Industry 4.0, semantic web services are indispensable to provide a reasonable abstraction of the underlying manufacturing capabilities. In this paper, we present semantic web services for AI-based research with physical factory simulation models in Industry 4.0. Therefore, we developed 68 semantic web services based on Web Ontology Language for Web Services (OWL-S) and Web Service Modeling Ontology (WSMO) and linked them to an already existing domain ontology for intelligent manufacturing control. Suitable for the requirements of CPS environments, our pre- and postconditions are verified in near real-time by invoking other semantic web services in contrast to complex reasoning within the knowledge base. Finally, we evaluate our implementation by executing a cyber-physical workflow composed of semantic web services using a state-of-the-art workflow management system.

@inproceedings{klein_complexdatagen_2019,
	author={Klein, Patrick and Bergmann, Ralph},
	title={{Generation of Complex data for AI-Based Predictive Maintenance Research With a Physical Factory Model}},
	booktitle={16th {International} {Conference} on {Informatics} in {Control} {Automation} and {Robotics}, {ICINCO} 2019, {Prague}, {Czech} {Republic}, {Proceedings} },
	year={2019},
	pages={40-50},
	volume = {1},
	publisher={SciTePress.},
	organization={INSTICC},
	doi={10.5220/0007830700400050},
	isbn={978-989-758-380-3},
	url={http://www.wi2.uni-trier.de/shared/publications/2019_KleinBergmann_Predictive_Maintenance_data_generation_machine_learning.pdf}
}

@inproceedings{klein_ftonto_2019,
	title = {{FTOnto: A Domain Ontology for a Fischertechnik Simulation Production Factory by Reusing Existing Ontologies}},
	author = {Klein, Patrick and Malburg, Lukas and Bergmann, Ralph},
	  editor    = {Robert J{\"{a}}schke and
               Matthias Weidlich},
	booktitle = {Proceedings of the Conference on "Lernen, Wissen, Daten, Analysen",
               Berlin, Germany, September 30 - October 2, 2019.},
	series    = {{CEUR} Workshop Proceedings},
	volume    = {2454},
	pages     = {253--264},
	publisher = {CEUR-WS.org},
	keywords = {{Ontology Engineering, Industry 4.0, Simulation Factory, Fischertechnik}},
	year      = {2019},
	abstract={Nowadays, semantic information provided by an ontology is indispensable in the context of Industry 4.0, especially when using methods from Artificial Intelligence. The currently available ontologies do not satisfy the demands of simulation environments used for research purposes. For this reason, we develop an ontology customized to Fischertechnik simulation factories by reusing existing ontologies. The ontology has been created according to requirements from two use cases. In our evaluation, it is determined that the ontology is suitable to represent machine components and their relationships while satisfying the specified requirements.},
	url = {http://www.wi2.uni-trier.de/shared/publications/2019_KleinMalburgBergmann_LWDA.pdf}
}

Nowadays, semantic information provided by an ontology is indispensable in the context of Industry 4.0, especially when using methods from Artificial Intelligence. The currently available ontologies do not satisfy the demands of simulation environments used for research purposes. For this reason, we develop an ontology customized to Fischertechnik simulation factories by reusing existing ontologies. The ontology has been created according to requirements from two use cases. In our evaluation, it is determined that the ontology is suitable to represent machine components and their relationships while satisfying the specified requirements.

@inproceedings{klein_data_generation_2018,
	publisher = {CEUR-WS.org},
	title = {{Data Generation with a Physical Model to Support Machine Learning Research for Predictive Maintenance}},
	url = {http://www.wi2.uni-trier.de/publications/2018_KleinBergmann_LWDA.pdf},
	booktitle = {Proceedings of the Conference "Lernen, Wissen, Daten, Analysen", {LWDA} 2018},
	author = {Klein, Patrick and Bergmann, Ralph},
	volume    = {2191},
	pages     = {179--190},
	year = {2018},
	keywords = {{Data Generation, Machine Learning, Predictive Maintenance, Industry 4.0}},
	abstract = {Today, manufacturing machines are continuously equippedwith various sensors, whose data enable to derive a comprehensive picture of the current state of each machine. Predictive maintenance approaches make use of  this data in order to predict the occurrence of possible failures before they actually occur, thereby significantly reducing production and service costs. The application of machine learning to sensor data streams is an essential part of data-driven predictive maintenance in order to find the patterns in  the data that are indicators of upcoming faults.  Thus, research on machine learning for  predictive maintenance is a recent and very challenging field. However, there are currently  no appropriate data sets available that can be used for thiskind of research. In this paper we therefore propose an approach for the generation of predictive maintenance data by using a physical Fischertechnik model factory equipped with several sensors. Different ways of reproducing real failures using this model are presented as well as a general procedure for data generation.}
}

Today, manufacturing machines are continuously equippedwith various sensors, whose data enable to derive a comprehensive picture of the current state of each machine. Predictive maintenance approaches make use of this data in order to predict the occurrence of possible failures before they actually occur, thereby significantly reducing production and service costs. The application of machine learning to sensor data streams is an essential part of data-driven predictive maintenance in order to find the patterns in the data that are indicators of upcoming faults. Thus, research on machine learning for predictive maintenance is a recent and very challenging field. However, there are currently no appropriate data sets available that can be used for thiskind of research. In this paper we therefore propose an approach for the generation of predictive maintenance data by using a physical Fischertechnik model factory equipped with several sensors. Different ways of reproducing real failures using this model are presented as well as a general procedure for data generation.