Trophallaxis is the mutual exchange and direct transfer of liquid food among eusocial insects such as ants, termites, wasps, and bees. This process allows efficient dissemination of nutrients and is crucial for the colony’s survival. In this paper, we present a data-driven agent-based model and use it to explore how the interactions of individual bees, following simple, local rules, affect the global food distribution. We design the rules in our model using laboratory experiments on honeybees. We validate its results via comparisons with the movement patterns in real bees. Using this model, we demonstrate that the efficiency of food distribution is affected by the density of the individuals, as well as the rules that govern their behavior: e.g., how they move and whether or not they aggregate. Specifically, food is distributed more efficiently when donor bees do not always feed their immediate neighbors, but instead prioritize longer motions, sharing their food with more-distant bees. This non-local pattern of food exchange enhances the overall probability that all of the bees, regardless of their position in the colony, will be fed efficiently. We also find that short-range attraction improves the efficiency of the food distribution in the simulations. Importantly, this model makes testable predictions about the effects of different bee densities, which can be validated in experiments. These findings can potentially contribute to the design of local rules for resource sharing in swarm robotic systems.

中文翻译：

---

蜜蜂群中数据驱动的资源分配建模

Trophallaxis是在诸如昆虫的蚂蚁，白蚁，黄蜂和蜜蜂等社会昆虫之间相互交换和直接转移液体食物。这个过程可以有效地传播营养，这对于殖民地的生存至关重要。在本文中，我们提出了一个基于数据驱动的基于代理的模型，并使用它来研究蜜蜂的相互作用如何遵循简单的本地规则影响全球食品分配。我们使用蜜蜂的实验室实验设计模型中的规则。我们通过与真实蜜蜂的运动模式进行比较来验证其结果。使用该模型，我们证明了食物分配的效率受个体密度以及支配其行为的规则的影响：例如，他们如何移动以及他们是否聚集。特别，当供体蜜蜂不总是喂饱其近邻时，而是优先考虑更长的运动，与更远处的蜜蜂分享食物，这样就可以更有效地分配食物。这种非本地的食物交换方式提高了所有蜜蜂（无论它们在殖民地中的地位如何）的有效饲养的总体可能性。我们还发现，在模拟中，短距离吸引可提高食物分配的效率。重要的是，此模型使 我们还发现，在模拟中，短距离吸引可提高食物分配的效率。重要的是，该模型使 我们还发现，在模拟中，短距离吸引可提高食物分配的效率。重要的是，此模型使关于不同蜜蜂密度影响的可预测的预测，可以在实验中得到验证。这些发现可能有助于在群体机器人系统中设计资源共享的本地规则。