Modelling Ecosystems with Deep Reinforcement Learning

Explore the innovative ecosystem simulator Ecotwin in this conference talk by Professor Claes Strannegård. Discover how three-dimensional terrain models based on real geographical data are combined with individual animal models controlled by deep reinforcement learning. Learn about the simulator's applications in studying both pristine ecosystems and those affected by human economic activities, including resource exploitation, land use change, pollution, and climate change. Gain insights into how this analytical tool can be utilized at the planning stage to understand local ecological consequences of various economic activities, promoting more eco-friendly decision-making. Presented by Claes Strannegård, a professor at the Department of Applied IT at the University of Gothenburg, this 27-minute talk was recorded at the 2023 GAIA Conference in Gothenburg, Sweden.