Control of Evolutionary Games on Networks

Abstract:

In this talk, I will discuss some recent work on the control of evolutionary games on networks, in which the nodes represent agents engaged in multiple simultaneous 2-player games and the edges define who plays with whom. The dynamics are governed by a deterministic update rule where after each round of games, agents adopt strategies based on the performance of their neighbors, while a subset of agents can be assigned strategies and thus serve as control inputs. We seek here solutions to a set of optimal control problems on network games, and in particular, the smallest set of control agents needed to drive the network to a desired strategy state. Since exact solution algorithms for this problem are computationally intensive, we look to take advantage of the structure of the networks. The main contributions of this work are (1) analytical solutions for networks with highly regular structure, (2) fast approximation algorithms for tree networks, and (3) an approach for approximating the solution on arbitrary networks by recursively decomposing the problem using graph partitioning.

Biography:http://www.rug.nl/staff/j.r.riehl/research?lang=en