Scholarlink

Swarm intelligence describes collective problem-solving that emerges from adaptive group behavior and decentralized control. This concept has led to a variety of bio-inspired optimization algorithms, many of which require further systematic experimental validation. In this study, we reexamine the biological validity of the BEECLUST algorithm by applying the experimental conditions and performance metrics of swarm robotics to a biological context. We find that real honeybees achieve aggregation at the thermal optimum even in relatively small groups. This indicates that the behavioral transition from individual to collective dynamics already emerges at low group sizes, suggesting that the most informative regime may occur at lower densities than reported in robotic implementations of the BEECLUST algorithm. We discuss possible explanations for this discrepancy and propose directions for future refinement.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Closing the Loop Between BEECLUST and Honeybee Thermotaxis

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Martin Stefanec,
  • Alexander Herlitz,
  • Daniel Reisinger,
  • Johannes Diebold,
  • Daniel Nicolas Hofstadler,
  • Anna Reichenpfader,
  • Laurenz Alexander Fedotoff,
  • Farshad Arvin,
  • Ronald Thenius,
  • Thomas Schmickl

摘要

Swarm intelligence describes collective problem-solving that emerges from adaptive group behavior and decentralized control. This concept has led to a variety of bio-inspired optimization algorithms, many of which require further systematic experimental validation. In this study, we reexamine the biological validity of the BEECLUST algorithm by applying the experimental conditions and performance metrics of swarm robotics to a biological context. We find that real honeybees achieve aggregation at the thermal optimum even in relatively small groups. This indicates that the behavioral transition from individual to collective dynamics already emerges at low group sizes, suggesting that the most informative regime may occur at lower densities than reported in robotic implementations of the BEECLUST algorithm. We discuss possible explanations for this discrepancy and propose directions for future refinement.