SOgym

Gym environment for developing structural optimization problems using deep reinforcement learning.

The environment is based on the topology optimization framework of Moving Morphable Components [1]. The design task is framed as a sequential decision process where at each timestep, the agent has to place one component.

Boundary Conditions

The environment samples from the following boundary conditions distribution at the start of each episode:

Table 1: Parameters Defining the Boundary Conditions Distribution

Parameter	Name	Distribution
h	Height	[1.0, 2.0]
w	Width	[1.0, 2.0]
L_s	Support Length	50% to 75%
P_s	Support Position	0 to (100% of L_s)
P_L	Load Position	0% to 100% of boundary opposite from support
θ_L	Load Orientation	[0°,360°] *

*The selected angle is filtered to ensure there is at least 45 degrees of difference with the support normal.

The blue wall represents a fully supported boundary and the red boundary the region where a unit load with varying orientation is randomly placed.

The environment's reward function can be modified to fit multiple constrained topology optimization objectives such as:

• Compliance minimization under hard volume constraint [Implemented]
• Compliance minimization under soft volume constraint [Implemented]
• Compliance minimization under global/local stress constraint
• Volume minimization under compliance constraint
• Combined volume and compliance minimization

SOgym Leaderboard

Observation Space Configurations and Algorithms

Observation Space	PPO	SAC	DreamerV3
Dense	Result for PPO	Result for SAC	Result for DreamerV3
Image	Result for PPO	Result for SAC	Result for DreamerV3
TopOpt Game	Result for PPO	Result for SAC	Result for DreamerV3

---

Citation

To cite this library, please refer to the following paper:

Rochefort-Beaudoin, T., Vadean, A., Aage, N., & Achiche, S. (2024). Structural Design Through Reinforcement Learning. arXiv preprint arXiv:2407.07288.

---

References

[1] Zhang, W., Yuan, J., Zhang, J. et al. A new topology optimization approach based on Moving Morphable Components (MMC) and the ersatz material model. Struct Multidisc Optim 53, 1243–1260 (2016). https://doi.org/10.1007/s00158-015-1372-3

[2] Nobel-Jørgensen, Morten & Malmgren-Hansen, David & Bærentzen, Andreas & Sigmund, Ole & Aage, Niels. (2016). Improving topology optimization intuition through games. Structural and Multidisciplinary Optimization. 54. 10.1007/s00158-016-1443-0.