Genetic Algorithm Learning of Nash Equilibrium

Abstract

To select or change a service provider, customers use the best compromise between price and quality of service (QoS). In this work, the authors formulate a game theoretic framework for the dynamical behaviors of Service Providers (SPs). They share the same market and are competing to attract more customers to gain more profit. Due to the divergence of SPs interests, it is believed that this situation is a non-cooperative game of price and QoS. The game converges to an equilibrium position known Nash Equilibrium. Using Genetic Algorithms (GAs), the authors find strategies that produce the most favorable profile for players. GAs are from optimization methods that have shown their great power in the learning area. Using these meta-heuristics, the authors find the price and QoS that maximize the profit for each SP and illustrate the corresponding strategy in Nash Equilibrium (NE). They also show the influence of some parameters of the problem on this equilibrium.