Decentralized inertial best-response with voluntary and limited communication in random communication networks

Multiple autonomous agents interact over a random communication network to maximize their individual utility functions which depend on the actions of other agents. We consider decentralized best-response with inertia type algorithms in form beliefs about future players based local information, and take that expected utilities computed respect these or continue previous actions. show convergence types Nash equilibrium weakly acyclic games. The result depends condition belief update information exchange protocols successfully learn positive probability finite time given static environment, i.e., when agents’ do not change. design fictitious play algorithm voluntary limited (DFP-VL) satisfy this condition. In protocol, each agent decides whom by assessing novelty its potential effect others’ beliefs. protocol entails sending only most frequent action they decide communicate with. Numerical experiments target assignment game demonstrate can more than halve number attempts while retaining same rate as DFP constantly attempt communicate.