Methods for Influencing a Flock

This page presents videos to complement our submitted Swarm Intelligence article.

Note that in each of our videos, the ad hoc agents are pink and the rest of the flock is grey. The grey box on the left shows the simulation parameters for the experiment.

Flock Behavior Without Influencing Agents

The following video shows the behavior of the flock when not being influenced by any influencing agents. In this case, each agent orients itself towards the average heading of its neighbors.

This video depicts 4 separate trials, where each trial is concluded when the flock has converged to traveling at a particular heading.

Influencing Agent Behaviors

In the following videos, we show the three main type of influencing agent behavior. Each video uses a random seed of 55 and a flock size of 100. The box below the grey simulation parameters box prints the number of steps needed for the flock to converge to the target direction (facing directly south) at the moment convergence occurs.

1-Step Lookahead Behavior

(58 steps for convergence)

1-Step Lookahead Behavior (58 steps for convergence)

2-Step Lookahead Behavior

(62 steps for convergence)

2-Step Lookahead Behavior (62 steps for convergence)

Coordinated Behavior

(66 steps for convergence)

Coordinated Behavior (66 steps for convergence)

Effect of Flock Size and Influencing Agent Percentage in Behavior Experiments

In the videos below, we show one example (where each video uses the same random seed for initial agent placement and orientation) of each of these variations. These videos show how these variations affect the dynamics of the agents in the environment. In each video, we use the 1-step lookahead algorithm. The box below the grey simulation parameters box prints the number of steps needed for the flock to converge to the target direction (facing directly south) at the moment convergence occurs.

Using the experimental setup described in Section 5.2

(flock size = 200, ad hoc agent percent = 10%)

Using the experimental setup described in Section 5.2 (flock size = 200, ad hoc agent percent = 10%)

Percentage of ad hoc agents in the flock decreased to 5%

(a smaller percentage of ad hoc agents in the flock can result in slower convergence)

Percentage of ad hoc agents in the flock decreased to 5% (a smaller percentage of ad hoc agents in the flock can result in slower convergence)

Percentage of ad hoc agents in the flock increased to 20%

(a larger percentage of ad hoc agents in the flock can result in faster convergence)

Percentage of ad hoc agents in the flock increased to 20% (a larger percentage of ad hoc agents in the flock can result in faster convergence)

Size of the flock decreased to 100

(a smaller flock is usually more spread-out initially, so convergence can occur slower since agents often have few neighbors)

Size of the flock decreased to 100 (a smaller flock is usually more spread-out initially, so convergence can occur slower since agents often have few neighbors)

Size of the flock increased to 300

(a larger flock is usually more compact initially, so convergence can occur quickly since agents often have many neighbors)

Size of the flock increased to 300 (a larger flock is usually more compact initially, so convergence can occur quickly since agents often have many neighbors)

Influencing Agent Placement Approaches

We consider four different methods in which to determine these initial positions at which to drop the influencing agents. Below we present videos of each of these methods. We show videos in which the number of agents (n) is 10, since this is one value of n that was used to obtain the data presented in Figure 11 in our paper.

As we present videos below, note that we use the same notation as in the paper. Specifically, the number of agents is represented by n, the number of influencing agents is represented by k, and the number of flocking agents is represented by m, such that k+m=n. For the videos shown below, we always use a random seed of 1 for calculating the initial flocking agent positions.

On the bottom right of the videos, you will notice that some numbers are printed out part way though each trial. These numbers are (from left to right): the number of time steps for the flock to converge, a number irrelevant to these trials (always 0.0), the number of flocking agents ’lost’, the x-axis average distance from each not ’lost’ flocking agent to the center of the flocking agents that are not ’lost’, and the y-axis average distance from each not ’lost’ flocking agent to the center of the flocking agents that are not ’lost'.