Türkiye from transit corridor to energy hub: A hybrid framework of an AI-based optimization system and market resilience

Türkiye stands at a critical juncture, aiming to transform its unique geography bridging Caspian and Middle Eastern resources with European markets from a transit corridor into a regional energy hub. This strategic ambition is accelerated by shifting EU energy policies post-2022, which prioritize supply diversification and route security. However, realizing this potential is constrained by infrastructure bottlenecks, import dependence risks, and rigid operational protocols. Existing literature typically divides these challenges into separate geopolitical or techno-economic analyses, failing to quantify the interplay between physical constraints and market resilience. This thesis addresses that gap by coupling a novel AI-driven Network Flow Optimization System (NFOS) enhanced by Deep Reinforcement Learning (Double DQN) with a Partial Equilibrium plus Monte Carlo (PEM+MC) market model. This hybrid framework couples AI-driven operational control with stochastic economic resilience assessment to simulate the Turkish gas system under complex stress scenarios. The results identify that AI-assisted dispatch reduced operating costs by 7% and improved flow efficiency by 17% compared to static baselines, successfully restoring 100% of service during severe supply cuts where traditional methods failed. Strategic analysis shows near-perfect resilience against single-source shocks but revealing systemic vulnerabilities under combined geopolitical and climatic stress. This study provides robust quantitative evidence that by leveraging Artificial Intelligence for network optimization and embracing flexible market mechanisms, Türkiye can transform its geographic destiny into a sustainable, economically resilient energy hub.