Call for Participation: DOCS 2026 EvoPINN Competition

The EvoPINN Competition focuses on large-scale multiobjective optimization for physics-informed neural network training. It aims to explore the intersection of evolutionary computation, black-box optimization, and scientific machine learning for PDE-solving tasks. Participants are invited to optimize PINN parameters on 12 representative EvoPINN benchmark problems under gradient-free and black-box evaluation settings, with final ranking based on Hypervolume.

We warmly welcome researchers, students, and practitioners interested in physics-informed neural networks, large-scale optimization, multiobjective optimization, evolutionary computation, scientific machine learning, and data-driven optimization of complex systems.