Feasibility-driven DDP solver reproduction and comparison based on crocoddyl
What happened
A researcher reproduced, isomorphically to the core mathematical formulation of the crocoddyl solver, the Riccati backward sweep, LM regularization, and gap contraction expansion, and conducted a three-way comparison on a cart-pole large swing-up task. Results show that FeasShoot's iteration count is less than 30% of that of the shooting-based method, and the regularization remains frozen. The above conclusions come from the comparison results of this reproduction experiment and have not yet been independently verified.
AI 根据报道生成 · 6 days ago更新
Reporting timeline
Follow the reporting to explore different perspectives.
- Guyuehomecrocoddyl: Feasibility-Driven Differential Dynamic Programming for Optimal Control
Reproduced the core mathematical formulation of the crocoddyl solver isomorphically: Riccati backward sweep, LM regularization, and gap contraction expansion. On a large-amplitude cart-pole swing-up task, a three-way comparison was conducted; FeasShoot required fewer than 30% of the iterations of the shooting method, and regularization remained frozen.
Activity for this event
- Current
- 3
- Peak
- 9Nov 4 22:00
- Past 24 hours
- -50%
The trend compares participants observed throughout the same period, so its range may be smaller than the current interest count. Move or tap to inspect each hour; use the arrow keys on a keyboard.