The impact of interaction radius on soft control performance based on the Vicsek Model
Caiyun Wang, Jing Han · 2016
The interaction neighborhood relationship among agents in many multi-agent systems is defined based on interaction radius r. Therefore, r plays an important role in the dynamics of these multi-agent systems. So does it also have influence on intervention methods for multi-agent systems? To answer this question, we study the impact of r on the performance of the Soft Control intervention method based on the simplified Vicsek Model. According to the soft control idea, some special agents, called shills, are added into the system with the purpose of changing the convergent heading of the group. Through simulations, we find that the interaction radius r has very different impact on the soft control performance in different cases: large r gives fast convergent speed in the centered fixed heading shills case, while small r leads to a large change in convergent heading in the centered evolvable heading shills case, and r has little impact on the performance for the distributed shills cases. This can be explained by the tradeoff of the positive and negative effects of large r. The `core' effect, `alliance' effect and `virtual shills' effect are discovered in the centered shills cases, which show the advantages of shills staying together.