1 - 10 of 11
-
PublicationEfficient Production Scheduling by Exploiting Repetitive Product Configurations( 2023)We consider the problem of scheduling production jobs on a single machine with sequence dependent family setup times and individual job deadlines. Given a set of jobs, the goal is to minimize the total time to process all jobs while every job meets its deadline. We study algorithms that compute an exact solution to the problem. Motivated by one example use case, we exploit a natural structural observation that occurs in many production settings: the number of product configurations may be significantly smaller than the total number of jobs.We identify an algorithm that is efficient in this setting in terms of performance. We experimentally evaluate its running time and compare it with two other natural approaches of exact job scheduling.
-
PublicationPotentials of Explainable Predictions of Order Picking Times in Industrial Production( 2023)The order picking process in a manufacturing supermarket is central in many industrial productions as it ensures that the items required for production are provided at the right time. However, the order picking process itself often is a black box, i.e., the time it takes to pick an order and the dependencies in the process that influence the time usually are not exactly known. In this work, we highlight the potentials of creating explainable predictions of order picking times using Artificial Intelligence methods. The prediction is based on the analysis of a historic database and on a linear regression analysis that learns the dependencies in the data. From this prediction, (1) the potential of identifying features having a high and a low influence on the order picking time, (2) the potential of optimizing the order picking process itself, and (3) the potential of optimizing depending processes are identified. For prediction, we utilize the regression methods LASSO and Decision Tree. These methods are compared with regard to their interpretability and usability in industrial manufacturing.
-
PublicationOptimization of a High Storage System with two Cranes per Aisle( 2023)Automated storage and retrieval systems (ASRS) are important in distribution centers and warehouses. To decrease cost or CO2 emissions it is natural to optimize various aspects of an ASRS. In this work, we provide a concept for a twophase optimization combining two important optimization tasks in ASRS: Given multiple rearrangement jobs, we first sequence these jobs to minimize the total travelling distance of the cranes. We continue the optimization by computing optimal trajectories for the sequence to guarantee energy efficient driving of the cranes. We describe our algorithms for a complex ASRS architecture with two cranes on parallel rails in one aisle. Additionally, we describe how to use our results for parallelization of crane movements in the considered warehouse architecture.
-
PublicationDeutsche Normungsroadmap Künstliche Intelligenz(DIN, 2022-12-09)
-
PublicationExplaining solutions to multi-stage stochastic optimization problems to decision makers( 2022)Decision support systems have become a critical component in the planning processes of companies needing to solve difficult optimization problems. Multi-stage, stochastic optimization problems pose a particular challenge for decision makers, as the uncertainty in the input data makes it hard to determine the correct decisions. The scalable stochastic optimization (SSO) technique proposes a way of solving these problems, but is not able to provide feedback to a decision maker regarding why it makes its decisions. We suggest a mechanism for explaining the feedback of SSO to help decision makers better understand a decision support system’s recommendations.
-
PublicationGenerating Causal Hypotheses for Explaining Black-Box Industrial Processes( 2022)Industrial processes nowadays are complex net-works of various modules. Due to their complexity, the processes are black-boxes, i.e. dependencies between input and output products of the process are mainly unknown. Getting insights into the process is challenging, since the extent of the data collected is usually to large to allow for a direct interpretation. Causal dependencies, i.e. the effect of an input variable on an output variable, allow for explaining what is happening inside the process. However, it is well known, that the identification of causal dependencies cannot be done purely data-driven but always requires at least a minimum of expert knowledge. In this article, a data-driven approach for the generation of causal hypotheses, i.e. dependencies in the data that might indicate a causal dependency, is presented. The approach is based on correlating input and output data of the black-box approach and quantifying the found correlation using regression. The approach is applied to an industrial compounding process and it is shown, that significant causal hypotheses can be found. These hypotheses are later validated by process experts.
-
PublicationKonzeptualisierung als Kernfrage des Maschinellen Lernens in der Produktion( 2020)Die Mehrheit der Projekte zur Überwachung und Diagnose Cyber-Physischer Systeme (CPS) beruht heute auf von Experten erstellten Modellen. Diese Modelle sind jedoch nur selten verfügbar, sind oft unvollständig, schwer zu überprüfen und zu warten. Datengetriebene Ansätze sind eine viel versprechende Alternative: Diese nutzen die großen Datenmengen die heutzutage in CPS gesammelt werden. Algorithmen verwenden die Daten, um die zur Überwachung notwendigen Modelle automatisch zu lernen. Dabei sind mehrere Herausforderungen zu bewältigen, wie zum Beispiel die Echtzeit-Datenerfassung und Speicherung, Datenanalyse, Mensch-Maschine Schnittstellen, Feedback- und Steuerungsmechanismen. In diesem Beitrag wird eine kognitive Referenzarchitektur vorgeschlagen, um diese Herausforderungen in Zukunft einfacher zu lösen. Anhand dieser Referenzarchitektur wird eine Schlüsselfrage diskutiert: Der Übergang von subsymbolische Informationen wie sie für das maschinelle Lernen typisch sind und symbolischen Informationen, welche von Menschen einfacher verstanden werden, d. h. die Frage der Konzeptualisierung. Anwendungsfälle aus unterschiedlichen Branchen werden schematisch dargestellt und untermauern die Richtigkeit und den Nutzen der Architektur.
-
PublicationMapping data sets to concepts using machine learning and a knowledge based approach( 2018)Machine learning techniques have a huge potential to take some tasks of humans, e.g. anomaly detection or predictive maintenance, and thus support operators of cyber physical systems (CPSs). One challenge is to communicate algorithms results to machines or humans, because they are on a sub-symbolical level and thus hard to interpret. To simplify the communication and thereby the usage of the results, they have to be transferred to a symbolic representation. Today, the transformation is typically static which does not satisfy the needs for fast changing CPSs and prohibit the usage of the full machine learning potential. This work introduces a knowledge based approach of an automatic mapping between the sub-symbolic results of algorithms and their symbolic representation. Clustering is used to detect groups of similar data points which are interpreted as concepts. The information of clusters are extracted and further classified with the help of an ontology which infers the current operational state. Data from wind turbines is used to evaluate the approach. The achieved results are promising, the system can identify its operational state without an explicit mapping.
-
PublicationIntegrating semantics for diagnosis of manufacturing systems( 2016)Trends in novel manufacturing systems lead to an increased level of data availability and smart usage of these data. Nowadays, many approaches are available to use the data, but because of an increased flexibility of the systems the interaction between machines and humans has become a challenge. Humans have to browse through a huge amount of data, need knowledge about the machine and underlying algorithms to interpret the results; they cannot use their known terms for communication, we call it the conceptual gap. The user should be enabled to communicate with the machine on a more abstract level and in a more natural way. Therefore, a natural language layer is introduced to provide users with a familiar interaction interface. Underlying layers contain knowledge about the domain, the machines and how data can be accessed and processed. This enables users' questions such as ""Are there any anomalies in the system?"" to be answered. Answers are provided in natural language and evaluated with a test set of 204 questions.
-
PublicationKognitive Architektur zum Konzeptlernen in technischen Systemen( 2015)Durch die Komplexität technischer Systeme und die benötigte Flexibilität für Systemänderungen werden innovative Ansätze benötigt die schnell an neue Situationen angepasst werden können. In dieser Arbeit wird die Implementierung der Architektur CATS (Cognitive Architecture for Technical Systems) beschrieben. Im Vergleich mit bestehen Architekturen existieren einige Unterschiede: Die Kommunikation zwischen System und Bediener wird über natürliche Sprache realisiert, wobei die Sprachverarbeitung auf dem automatischen Lernen von Konzepten besteht. Da technische Systeme ihr Verhalten über die Zeit ändern, kann CATS nicht auf statische Textdokumente als Wissensbasis zurückgreifen. Stattdessen müssen die nötigen Informationen aus Echtzeitdaten extrahiert werden. Um die Funktion von CATS zu validieren, wurde die Architektur auf einem Demonstrator implementiert. Das Stanford CoreNLP Framework wurde darin zur Sprachverarbeitung benutzt. Die Wissensmodellierung wird durch die Web Ontology Language (OWL) realisiert und die Vorverarbeitung der Maschinendaten mit verschiedenen maschinellen Lernalgorithmen implementiert. Der Demonstrator ist in der Lage 84% der Eingabetexte richtig zu beantworten. Durch die Benutzung von CATS in technischen Systemen kann der Installations- und Adaptionsaufwand reduziert werden, da Ansätze für maschinelles Lernen Informationen über das System automatisch gewinnen. Weiterhin wird der Informationsaustausch zwischen Bediener und Maschine durch natürliche Sprache und Konzepte vereinfacht.
