Duan Haibin, Sun Changhao
1] State Key Laboratory of Virtual Reality Technology and Systems, Beihang University, Beijing 100191, P. R. China [2] Science and Technology on Aircraft Control Laboratory, School of Automation Science and Electronic Engineering, Beihang University, Beijing 100191, P. R. China.
Sci Rep. 2014 Jun 9;4:5210. doi: 10.1038/srep05210.
Many hostile scenarios exist in real-life situations, where cooperation is disfavored and the collective behavior needs intervention for system efficiency improvement. Towards this end, the framework of soft control provides a powerful tool by introducing controllable agents called shills, who are allowed to follow well-designed updating rules for varying missions. Inspired by swarm intelligence emerging from flocks of birds, we explore here the dependence of the evolution of cooperation on soft control by an evolutionary iterated prisoner's dilemma (IPD) game staged on square lattices, where the shills adopt a particle swarm optimization (PSO) mechanism for strategy updating. We demonstrate that not only can cooperation be promoted by shills effectively seeking for potentially better strategies and spreading them to others, but also the frequency of cooperation could be arbitrarily controlled by choosing appropriate parameter settings. Moreover, we show that adding more shills does not contribute to further cooperation promotion, while assigning higher weights to the collective knowledge for strategy updating proves a efficient way to induce cooperative behavior. Our research provides insights into cooperation evolution in the presence of PSO-inspired shills and we hope it will be inspirational for future studies focusing on swarm intelligence based soft control.
在现实生活中存在许多不利的情况，其中合作不受青睐，集体行为需要干预以提高系统效率。为此，软控制框架通过引入被称为托儿的可控智能体提供了一个强大的工具，这些托儿被允许遵循精心设计的更新规则来完成不同的任务。受鸟群中出现的群体智能启发，我们在此通过在方形晶格上进行的进化迭代囚徒困境（IPD）博弈来探索合作进化对软控制的依赖性，其中托儿采用粒子群优化（PSO）机制进行策略更新。我们证明，托儿不仅可以通过有效地寻找潜在的更好策略并将其传播给其他智能体来促进合作，而且通过选择合适的参数设置，可以任意控制合作的频率。此外，我们表明增加更多托儿并不能进一步促进合作，而在策略更新中为集体知识赋予更高权重被证明是诱导合作行为的有效方法。我们的研究为存在受PSO启发的托儿情况下的合作进化提供了见解，我们希望它将为未来专注于基于群体智能的软控制的研究提供启发。
