Limits...
Modelling group navigation: transitive social structures improve navigational performance.

Flack A, Biro D, Guilford T, Freeman R - J R Soc Interface (2015)

Bottom Line: Our results show that groups consisting of equally informed individuals achieve the highest level of accuracy when they are hierarchically organized with the minimum number of preferred connections per individual.More specifically, group navigation does not only depend on the underlying social relationships, but also on how much weight leading individuals put on following others.The results have broader implications for studies on collective navigation and motion because they show that only by considering a group's social system can we fully elucidate the dynamics and advantages of joint movements.

View Article: PubMed Central - PubMed

Affiliation: Department of Zoology, University of Oxford, South Parks Road, Oxford OX1 3PS, UK aflack@orn.mpg.de.

No MeSH data available.


Related in: MedlinePlus

Difference in navigational accuracy between groups with (hierarchical and random) and without underlying network structure (mean ± s.e.m.) as a function of (a) average out-degree (w = 0.3) and (b) weighting factor (out-degree = 0.9). (c) Average distance to the centre of the flock (mean ± s.e.m.) as a function of the number of followers for hierarchical and random networks. Navigational accuracy is calculated as the distance of the group's centre of mass to the target at the end of the simulation. Groups with random or hierarchical networks are shown as light grey diamonds and dark grey circles, respectively. Inset shows the percentage of fragmented groups (i.e. those in which not all individuals remained within a distance of rA from their nearest neighbour) as a function of the weighting factor.
© Copyright Policy - open-access
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4528586&req=5

RSIF20150213F2: Difference in navigational accuracy between groups with (hierarchical and random) and without underlying network structure (mean ± s.e.m.) as a function of (a) average out-degree (w = 0.3) and (b) weighting factor (out-degree = 0.9). (c) Average distance to the centre of the flock (mean ± s.e.m.) as a function of the number of followers for hierarchical and random networks. Navigational accuracy is calculated as the distance of the group's centre of mass to the target at the end of the simulation. Groups with random or hierarchical networks are shown as light grey diamonds and dark grey circles, respectively. Inset shows the percentage of fragmented groups (i.e. those in which not all individuals remained within a distance of rA from their nearest neighbour) as a function of the weighting factor.

Mentions: Because networks may vary in the number of strong connections, we first explored how navigational performance changes as a function of its underlying social structure and the degree of connectedness within the group. We increased the number of strong connections from one per group (out-degree = 0.1) to one per individual (0.9) until reaching fully connected groups (4.6 for hierarchical groups). The change in navigational performance of hierarchical groups showed two phases. First, it increased with the number of strong connections until all individuals had one strong connection (0.9). Groups with random social structure did not show this pattern. Whereas hierarchical groups with an out-degree of 0.9 arrived on average 157.3 m (±340 m s.d.) closer to the target than groups with no network structure, random groups improved by only 41.2 ± 635 m (mean ± s.d.). Second, navigational performance of hierarchical groups decreased again after increasing the number of strong connections to more than one per individual (figure 2a). Again, this pattern was not observed for random groups. Therefore, to explore the largest possible difference in navigational accuracy between groups with random and directed networks, we focus in all remaining simulations on networks with an average out-degree of 0.9.Figure 2.


Modelling group navigation: transitive social structures improve navigational performance.

Flack A, Biro D, Guilford T, Freeman R - J R Soc Interface (2015)

Difference in navigational accuracy between groups with (hierarchical and random) and without underlying network structure (mean ± s.e.m.) as a function of (a) average out-degree (w = 0.3) and (b) weighting factor (out-degree = 0.9). (c) Average distance to the centre of the flock (mean ± s.e.m.) as a function of the number of followers for hierarchical and random networks. Navigational accuracy is calculated as the distance of the group's centre of mass to the target at the end of the simulation. Groups with random or hierarchical networks are shown as light grey diamonds and dark grey circles, respectively. Inset shows the percentage of fragmented groups (i.e. those in which not all individuals remained within a distance of rA from their nearest neighbour) as a function of the weighting factor.
© Copyright Policy - open-access
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4528586&req=5

RSIF20150213F2: Difference in navigational accuracy between groups with (hierarchical and random) and without underlying network structure (mean ± s.e.m.) as a function of (a) average out-degree (w = 0.3) and (b) weighting factor (out-degree = 0.9). (c) Average distance to the centre of the flock (mean ± s.e.m.) as a function of the number of followers for hierarchical and random networks. Navigational accuracy is calculated as the distance of the group's centre of mass to the target at the end of the simulation. Groups with random or hierarchical networks are shown as light grey diamonds and dark grey circles, respectively. Inset shows the percentage of fragmented groups (i.e. those in which not all individuals remained within a distance of rA from their nearest neighbour) as a function of the weighting factor.
Mentions: Because networks may vary in the number of strong connections, we first explored how navigational performance changes as a function of its underlying social structure and the degree of connectedness within the group. We increased the number of strong connections from one per group (out-degree = 0.1) to one per individual (0.9) until reaching fully connected groups (4.6 for hierarchical groups). The change in navigational performance of hierarchical groups showed two phases. First, it increased with the number of strong connections until all individuals had one strong connection (0.9). Groups with random social structure did not show this pattern. Whereas hierarchical groups with an out-degree of 0.9 arrived on average 157.3 m (±340 m s.d.) closer to the target than groups with no network structure, random groups improved by only 41.2 ± 635 m (mean ± s.d.). Second, navigational performance of hierarchical groups decreased again after increasing the number of strong connections to more than one per individual (figure 2a). Again, this pattern was not observed for random groups. Therefore, to explore the largest possible difference in navigational accuracy between groups with random and directed networks, we focus in all remaining simulations on networks with an average out-degree of 0.9.Figure 2.

Bottom Line: Our results show that groups consisting of equally informed individuals achieve the highest level of accuracy when they are hierarchically organized with the minimum number of preferred connections per individual.More specifically, group navigation does not only depend on the underlying social relationships, but also on how much weight leading individuals put on following others.The results have broader implications for studies on collective navigation and motion because they show that only by considering a group's social system can we fully elucidate the dynamics and advantages of joint movements.

View Article: PubMed Central - PubMed

Affiliation: Department of Zoology, University of Oxford, South Parks Road, Oxford OX1 3PS, UK aflack@orn.mpg.de.

No MeSH data available.


Related in: MedlinePlus