Generation and analysis of multiple futures with swarming agents

  • Authors:
  • H. Van Dyke Parunak

  • Affiliations:
  • Vector Research Center of TTGSI, Ann Arbor, MI

  • Venue:
  • Proceedings of the 9th International Conference on Autonomous Agents and Multiagent Systems: volume 1 - Volume 1
  • Year:
  • 2010

Quantified Score

Hi-index 0.00

Visualization

Abstract

Most agent-based modeling techniques generate one trajectory per run, greatly under-sampling the space of possible trajectories. Swarming agents that interact through digital pheromones can explore many alternative futures in parallel, based on an interpretation of pheromone fields as probability fields that yields more information from them than swarming models usually yield, and also facilitates integration with probability-based AI mechanisms.