On May 1, 2025, researchers Ahmed R. Sadik, Muhammad Ashfaq, Niko Mäkitalo, and Tommi Mikkonen released a pivotal study titled ´Urban Air Mobility as a System of Systems: An LLM-Enhanced Holonic Approach.´ The work details a novel integration of Large Language Models with holonic system architecture designed to address key challenges in urban air mobility (UAM), such as scalability, system adaptability, and complex resource management.
UAM operates in an environment marked by rapidly shifting variables—including weather, traffic, and airspace restrictions. The proposed intelligent holonic architecture leverages resource holons as semi-autonomous units coordinating air taxis, ground transport, and vertiports. By embedding Large Language Models within these holons, individual components gain context-awareness, advanced reasoning abilities, and improved inter-component communication. The approach moves decision-making away from rigid centralization, empowering real-time replanning, resource allocation, and disruption mitigation during events such as unexpected weather or airspace closures. A detailed case study involving electric scooters and air taxis demonstrated practical viability, showing how systems dynamically reallocate resources while maintaining overall network efficiency.
Holonic architectures fundamentally reimagine how systems of systems interact and evolve. Drawing inspiration from biological holons—self-contained units capable of independent or collective action—these frameworks are particularly adept at handling dynamism and unpredictability in complex domains such as disaster response, smart cities, healthcare logistics, and industrial automation. The integration of Large Language Models further enhances adaptability, scalability, and efficiency, as holons are now able to absorb real-time data, predict downstream effects, and autonomously coordinate with peer units. For example, LLM-powered holons can flexibly reroute urban vehicle traffic based on live congestion data or prioritize emergency response through intelligent context analysis.
The decentralized, Artificial Intelligence-driven holonic approach reduces bottlenecks, promotes rapid scaling, and improves resource usage across transportation networks. However, this transformation introduces new challenges. As control becomes distributed, concerns around cybersecurity, ethical transparency, and accountability rise, as does the need for robust scalability frameworks. Nonetheless, the research demonstrates significant promise across domains—ushering urban mobility and related fields into a new era of resilient, human-centric, and efficient ecosystems.
