Logo des Repositoriums

Auflistung nach Schlagwort "Artificial neural networks"

1 - 3 von 3
  • Zeitschriftenartikel
    Decision Support for the Automotive Industry
    (Business & Information Systems Engineering: Vol. 61, No. 4, 2019) Gleue, Christoph; Eilers, Dennis; Mettenheim, Hans-Jörg; Breitner, Michael H.
    In the automotive industry, it is very common for new vehicles to be leased rather than sold. This implies forecasting an accurate residual value for the vehicles, which is a major factor for determining monthly leasing rates. Either a systematic overestimation or underestimation of future residual values can incur large potential losses in resale value or, respectively, competitive disadvantages. For the purpose of facilitating residual value related management decisions, an operative decision support system is introduced with emphasis on its forecasting capabilities. In the paper, the use of artificial neural networks for this application is demonstrated in a case study based on more than 250,000 data sets of leasing contracts from a major German car manufacturer, completed between 2011 and 2017. The importance of determining price factors and the effect of different time horizons on forecasting accuracy are investigated and practical implications are discussed. In addition, the authors neither found a significant explanatory nor predictive power of external economic factors, which underlines the importance of collecting and taking advantage of vehicle-specific data or, in more general terms, the exclusive data of corporations, which is often only available internally.
  • Zeitschriftenartikel
    Lastmanagement in Stromnetzen
    (Wirtschaftsinformatik: Vol. 55, No. 1, 2013) Köpp, Cornelius; Mettenheim, Hans-Jörg; Breitner, Michael H.
    Die zunehmende Verbreitung dezentral eingespeister erneuerbarer Energien führt zu Stabilitätsproblemen in Stromnetzen. Im Gegensatz zu steuerbaren und somit gut planbaren konventionellen Kraftwerken ist die Energieerzeugung aus Windkraft und Photovoltaikanlagen aufgrund von wetterbedingten Fluktuationen nur deutlich kurzfristiger und ungenauer zu prognostizieren. Grundsätzlich können sowohl Stromerzeuger als auch -verbraucher helfen, die Stabilitätsproblematik zu verringern. Zu diesem Zweck werden Beiträge zu einem Entscheidungsunterstützungssystem mit Steuerungsstrategien zur Beeinflussung des Verhaltens von Stromerzeugern und -verbrauchern vorgeschlagen. Auf der Erzeugerseite dienen dezentrale, kleinere Blockheizkraftwerke im Verbund als virtuelles Kraftwerk zum Lastausgleich. Basierend auf Prognosen bietet der Betreiber des virtuellen Kraftwerks einen Lastgang an. Auf Abweichungen kann später durch geeignete Steuerung der Blockheizkraftwerke reagiert werden. Auf der Verbraucherseite kommt eine anreizbasierte Steuerung über Preissignale zum Einsatz. Intelligente Geräte reagieren auf die übermittelten Strompreise dadurch, dass sie ihre Ausführungszeit, wenn möglich, auf einen Zeitraum günstiger Strompreise verlegen. Zur Vorbereitung eines Feldversuchs ist es zunächst notwendig, die intelligenten Geräte zu simulieren. Das Ergebnis der Simulation wird dazu verwendet, ein künstliches neuronales Netz zu trainieren, mit dem sich für gewünschte Lastverschiebungen geeignete Preissignale ermitteln lassen.AbstractDecentralized renewable energy sources become more and more common. This leads to stability problems in power grids. Conventional energy sources are easy to control. In contrast, wind and solar power are much more difficult to forecast. Forecasts are only possible short term and are more imprecise. Producers and consumers of energy can try to help reducing stability problems. Contributions towards a decision support system are proposed and recommend how to alter the behavior of producers and consumers. On the producer side centrally controlled heat and power plants are able to shift load in a virtual power plant. The plant operator offers a load curve based on forecasts. The centrally controlled heat and power plants help to mitigate the effect of revised forecasts. An incentive based control on the consumer side is also proposed. Smart appliances react to pricing information. They alter their execution window towards the cheapest time slot, if possible. The exact behavior of appliances in the expected field experiment is still partially unknown. It is necessary to simulate the behavior of these appliances and to train an artificial neural network. The artificial neural network allows computing the pricing signal leading to a desired load shift.
  • Zeitschriftenartikel
    Towards Strong AI
    (KI - Künstliche Intelligenz: Vol. 35, No. 1, 2021) Butz, Martin V.
    Strong AI—artificial intelligence that is in all respects at least as intelligent as humans—is still out of reach. Current AI lacks common sense, that is, it is not able to infer, understand, or explain the hidden processes, forces, and causes behind data. Main stream machine learning research on deep artificial neural networks (ANNs) may even be characterized as being behavioristic. In contrast, various sources of evidence from cognitive science suggest that human brains engage in the active development of compositional generative predictive models (CGPMs) from their self-generated sensorimotor experiences. Guided by evolutionarily-shaped inductive learning and information processing biases, they exhibit the tendency to organize the gathered experiences into event-predictive encodings. Meanwhile, they infer and optimize behavior and attention by means of both epistemic- and homeostasis-oriented drives. I argue that AI research should set a stronger focus on learning CGPMs of the hidden causes that lead to the registered observations. Endowed with suitable information-processing biases, AI may develop that will be able to explain the reality it is confronted with, reason about it, and find adaptive solutions, making it Strong AI. Seeing that such Strong AI can be equipped with a mental capacity and computational resources that exceed those of humans, the resulting system may have the potential to guide our knowledge, technology, and policies into sustainable directions. Clearly, though, Strong AI may also be used to manipulate us even more. Thus, it will be on us to put good, far-reaching and long-term, homeostasis-oriented purpose into these machines.