Logo des Repositoriums

Auflistung nach Schlagwort "simulations"

1 - 2 von 2
  • Konferenzbeitrag
    Researching Financial Market Dynamics through Algorithmic Trading Agents
    (AKWI Jahrestagung 2024, 2024) Helstab, Paul; Mosler, Christof; Kaiser, Daniel
    The evolving landscape of financial markets demands the integration of advanced technologies and quantamental paradigms like market simulations and artificial intelligence. This study explores the use of agent-based interactive discrete event trading simulations to analyze financial market scenarios and detect anomalies. The research involves various market fees and synthetic competition between exchanges, utilizing an inter-market spread arbitrage machine for price stability. This approach enables the observation of anomalies related to spreads, execution speed, traded volumes, liquidity, order execution probability, and agent decision-making influence. This study could serve as a valuable tool for financial institutions and regulatory authorities in strategic decision-making concerning financial market challenges.
  • Conference demo
    Web-based prototype of a visual and interactive deep learning simulation
    (Proceedings of DELFI 2024, 2024) Koch, Christian; Salmen, Frederic; Schroeder, Ulrik
    Due to its rapid development and growing relevance in society, the research field of artificial intelligence (AI) is finding its way into more and more schools. One subsection of artificial intelligence, deep learning, is especially relevant since most of the rapid development of the study field in recent years can be attributed to it. Because of its complex nature, deep learning is very suitable to be taught in a visual and interactive way, which can benefit learning outcomes. Existing solutions can convey basic knowledge about deep learning, but none of them are well suited to customization or facilitating deep understanding. We introduce a visual and interactive deep learning simulation with rich possibilities for personalization by teachers and students, usable in any web context.