![[2602.17783] Multi-material Multi-physics Topology Optimization with Physics-informed Gaussian Process Priors](https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png){kind=logo}
[2602.17783] Multi-material Multi-physics Topology Optimization with Physics-informed Gaussian Process Priors
Summary
This paper presents a novel framework for multi-material, multi-physics topology optimization using physics-informed Gaussian processes, addressing challenges in existing machine learning approaches.
Why It Matters
The study addresses significant limitations in current topology optimization methods, particularly in handling complex physics and multi-material scenarios. By leveraging physics-informed Gaussian processes, the framework enhances computational efficiency and accuracy, which is crucial for advancements in engineering and materials science.
Key Takeaways
- Introduces a framework that combines physics-informed Gaussian processes with topology optimization.
- Addresses limitations of existing ML approaches in multi-material, multi-physics problems.
- Demonstrates effectiveness on benchmark problems, improving training speed and accuracy.
- Validates results using both open-source codes and commercial software.
- Offers insights into generating interpretable material distributions in design.
Computer Science > Machine Learning arXiv:2602.17783 (cs) [Submitted on 19 Feb 2026] Title:Multi-material Multi-physics Topology Optimization with Physics-informed Gaussian Process Priors Authors:Xiangyu Sun, Shirin Hosseinmardi, Amin Yousefpour, Ramin Bostanabad View a PDF of the paper titled Multi-material Multi-physics Topology Optimization with Physics-informed Gaussian Process Priors, by Xiangyu Sun and 3 other authors View PDF HTML (experimental) Abstract:Machine learning (ML) has been increasingly used for topology optimization (TO). However, most existing ML-based approaches focus on simplified benchmark problems due to their high computational cost, spectral bias, and difficulty in handling complex physics. These limitations become more pronounced in multi-material, multi-physics problems whose objective or constraint functions are not self-adjoint. To address these challenges, we propose a framework based on physics-informed Gaussian processes (PIGPs). In our approach, the primary, adjoint, and design variables are represented by independent GP priors whose mean functions are parametrized via neural networks whose architectures are particularly beneficial for surrogate modeling of PDE solutions. We estimate all parameters of our model simultaneously by minimizing a loss that is based on the objective function, multi-physics potential energy functionals, and design-constraints. We demonstrate the capability of the proposed framework on benchmark TO problems such a...
