Distributed server selection in Mobile Edge Computing through Game theory

This diploma project investigates distributed server selection in Mobile Edge Computing (MEC) through the lens of game theory. It systematically analyzes how various game-theoretic models---including Static Games of Complete Information, Congestion Games, and Bayesian Games---can optimize server selection while accounting for performance metrics like latency, energy consumption, and fairness. A core contribution of the project is the development and simulation of a Congestion Game model, which dynamically penalizes overloaded servers and improves load distribution. Simulations demonstrate how increasing user count or congestion sensitivity impacts cost trends and fairness, highlighting the model's scalability and adaptability. In contrast, the Static Game serves as a baseline for understanding equilibrium behavior under full information but lacks responsiveness to crowding. Results show that congestion-aware models outperform traditional static approaches in balancing efficiency and fairness, especially in high-density MEC systems. Overall, the thesis provides a practical, simulation-driven framework for applying strategic interaction models to real-world MEC challenges, offering insights into how game-theoretic models can support decentralized, scalable, and responsive decision-making.