cant understand how to scale the penalty coefficient in ”trajopt_optimizers/trajopt_sqp“.

tesseract-robotics / trajopt

Trajectory Optimization Motion Planner for ROS

379 stars 102 forks source link

cant understand how to scale the penalty coefficient in ”trajopt_optimizers/trajopt_sqp“. #353

Closed sailuluo closed 9 months ago

sailuluo commented 9 months ago

in ”trajopt_optimizers\trajopt_sqp\src\trust_region_sqp_solver.cpp“， It seems that it is necessary to update the constraint coefficient at the end of each loop through "TrustRegionSQPSolver::adjustPenalty()", But I can only find updates of the SQP result，cant find updates of qp_problem.

Levi-Armstrong commented 9 months ago

This is a know issue, I believe this is one the final remaining items to tackle to where the Ifopt version aligns with the legacy. Unfortunately, I have not had time to look into how best solve this. If you have any thoughts please let me know.

Levi-Armstrong commented 9 months ago

Addressed by #366