Modeling of
Agents
Reality is made up of interconnected individuals. Using agent-based modeling, at CAOS we simulate complex systems in which multiple autonomous software entities (agents) interact with one another. This approach allows us to study emergent collective behaviors, design intelligent social platforms, and create distributed solutions for human assistance. Some examples of this include Multi-Agent Systems (MAS) or, more directly, the simulation of virtual environments through the combination of agents.
Sistemas Multi-Agente
Software architectures in which agents cooperate or compete to solve decentralized problems
Simulación de Entornos Virtuales
Behavior Modeling and User Activity Recognition for Telecare or Assisted Training
Análisis de Flujo de Datos
Traffic flow analysis, urban land-use planning, disease spread simulation, and automated decision-making
Laboratory Milestones
Modeling Human Behavior: Behavioral models have been developed within the autonomous vehicle ecosystem in which, rather than treating pedestrians or other vehicles as mere static or predictable obstacles, they are modeled as interactive autonomous agents with intentions and internal states (fatigue, distraction, probable trajectories).
Cooperative Urban Mobility: The development of smart city simulation environments where diverse agents coexist: connected vehicles, smart traffic lights, traffic sensors, and control centers. Through this modeling, they have demonstrated how the exchange of information among agents can alleviate traffic congestion and reduce pollutant emissions.
Multi-Agent Logistics Fleets: Design of negotiation and consensus protocols in which each vehicle acts as an independent agent. When faced with an unforeseen event (such as a traffic jam or a breakdown), the agents “negotiate” with one another autonomously to redistribute tasks and replan optimal routes in real time, without the need for a central server to recalculate the entire problem.
European Success Story – Trainutri Project: The CAOS group led the development of the Trainutri social platform for seniors, funded by the European Union (Ambient Assisted Living Joint Program). The system used distributed intelligence and software agents to monitor, motivate, and guide the exercise and nutrition of older adults, substantially improving their quality of life.
«The most dangerous power is that of someone who commands but does not govern.»
– Gonzalo Torrente –