1: \begin{abstract}

2: We study the problem of asymptotic consensus as it occurs in a wide range of applications

3:   in both man-made and natural systems.

4: In particular, we study systems with directed communication graphs that may change over time.

5: 

6: We recently proposed a new family of convex combination algorithms in dimension one whose weights depend on the

7:   received values and not only on the communication topology.

8: Here, we extend this approach to arbitrarily high dimensions by introducing two new algorithms:

9:   the ExtremePoint and the Centroid algorithm.

10: Contrary to classical convex combination algorithms, both

11:   have component-wise contraction rates that are constant in the number of agents.

12: Paired with a speed-up technique for convex combination algorithms, we get a convergence time linear

13:   in the number of agents, which is optimal.

14: 

15: Besides their respective contraction rates, the two algorithms differ in the fact that 

16:   the Centroid algorithm's update rule is independent of any coordinate system 

17:   while the ExtremePoint algorithm implicitly assumes a common agreed-upon coordinate system

18:   among agents.

19: The latter assumption may be realistic in some man-made multi-agent systems but

20:   is highly questionable in systems designed for the modelization of natural phenomena.

21: 

22: Finally we prove that our new algorithms also

23:   achieve asymptotic consensus under very weak connectivity assumptions,

24:   provided that agent interactions are bidirectional.

25: 

26: \end{abstract}

27: