MUMBAI, India, Aug. 17 -- Intellectual Property India has published a patent application (202641094165 A) filed by Vardhaman College Of Engineering on August 04, 2026, for Ai-Driven Procedural World Generation Using Reinforcement Learning In Open-World Games.
Inventors include Mr. K Venkatesh; Mr. Kottu Santosh Kumar; Mr. Nirmal Keshari Swain; Mr. Shobanbabu R Jaganathan; Ms. Shruti S Soma; Mr. Pichala Dastagiri Reddy; and Dr. Ramesh M.
The application for the patent was published on August 14, 2026, under issue no. 33/2026.
Abstract: AI-Driven Procedural World Generation Using Reinforcement Learning in Open-World Games is the proposed invention. The invention disclosed herein proposes an artificial intelligence-driven procedural world generation system for open-world games using reinforcement learning and graph based environmental optimisation. It uses diffusion-based terrain synthesis to generate realistic terrains and a dynamic spatial graph to represent the changing world. A Graph Neural Network evaluates structural coherence, connectivity, ecological consistency, and navigational quality. A Proximal Policy Optimisation (PPO) reinforcement learning agent progressively generates terrains, settlements, ecosystems, resources, quests, and interactive environments by maximising a multi-objective reward function. It also uses player behaviour analysis to guide future world expansions based on exploration patterns, gameplay preferences, and mission interactions. A hierarchical chunk generation mechanism allows for scalable real-time world expansion with low computational overhead. The proposed invention creates immersive, adaptive, and diverse virtual worlds with improved realism, balanced gameplay, enhanced player engagement, and significantly reduced manual development effort compared to conventional procedural content generation methods.
Disclaimer: Curated by HT Syndication.
