Akademik Veri Yönetim Sistemi
ADVANCED THEORY AND SIMULATIONS, cilt.8, sa.9, 2025 (SCI-Expanded)
This study uses deep Q-learning reinforcement learning to introduce a novel optimization framework for semi-transparent organic solar cells (ST-OSCs). This approach integrates the Transfer Matrix Method with artificial intelligence to streamline design processes by addressing the dual challenge of achieving high visible light transparency and efficient photovoltaic performance. The research focuses on PBDB-T:ITIC-based active layers coupled with asymmetrical dielectric-metal-dielectric transparent contacts, optimizing layer thicknesses and material properties. The deep Q-learning algorithm efficiently navigates the complex design space, achieving a maximum average visible transmittance of 48.97% while maintaining strong photo-current density. This optimization balances transparency and absorption, which are critical for ST-OSCs, by reducing reflection losses and enhancing photon management. The study demonstrates the effectiveness of reinforcement learning in handling intricate multi-layer ST-OSCs, surpassing traditional optimization techniques. Results highlight the potential of adaptive learning algorithms in identifying high-performance material configurations, minimizing computational costs, and ensuring precision. This work underscores the transformative role of AI in renewable energy technologies, offering scalable, sustainable solutions to modern energy challenges. By advancing the integration of artificial intelligence and material science, this study opens new pathways for optimizing renewable energy technologies and improving the performance of optoelectronic devices.