search

robotology / osqp-eigen

Simple Eigen-C++ wrapper for OSQP library
https://robotology.github.io/osqp-eigen/
BSD 3-Clause "New" or "Revised" License
385 stars 114 forks source link

Why the solver returns "the problem is non-convex" or " solve successfull" #107

Open tmacattank opened 2 years ago

tmacattank commented 2 years ago

I am now using quadratic programming to solve a trajectory planning problem, and the test scenario is to overtake a vehicle. I try OSQP diresctly and OSQP eigen. OSQP works well, while when i use osqp eigen, sometimes it works well, sometimes it suddenly return "the problem is non-convex" or "solve unccessful". I check the error code from OSQP, and only when the Hessian matrix is not positive definite, can the problem becomes non-convex. I have follow the OSQP eigen MPC example grammar rules to use this tool, is there any suggestions on the randomly occured crash?

GiulioRomualdi commented 2 years ago

Hi @tmacattank. Could you please give us a minimum example that we can use to better debug the problem?

tmacattank commented 2 years ago

I am using an online simulation platform based on ROS to develop the algorithm, I will try to find an easy way so that you can run my code

YuRan-W commented 1 year ago

Hello, I also encountered this problem, did you solve it?