Approximate Run time of the code #1

saxenam06 · 2022-04-28T12:36:28Z

Hello, Many Thanks for your repository. I was trying to run merging_scenario_paper.py code. It displayed the following output in the console and then seems to have stucked since 4 hrs now. Does the code really takes time or something has gone wrong in my installation? Many Thanks for your help.

added constraint between 0 and 1 for 0
added constraint between 1 and 2 for 0
added constraint between 1 and 3 for 0
added constraint between 0 and 2 for 0
added constraint between 0 and 3 for 0
added constraint between 1 and 0 for 1
added constraint between 0 and 2 for 1
added constraint between 0 and 3 for 1
added constraint between 1 and 2 for 1
added constraint between 1 and 3 for 1

brechtevens · 2022-04-29T09:25:13Z

Hello,

Glad to hear you are trying out the simulator! The game starts in a 'paused' state, you just need to press the spacebar to start the game. I just added the instructions for interfacing with the simulator in the readme. I also packaged the src code now using flit, which is better practice than the weird implicit imports I did before.

Let me know how it goes!

Best,
Brecht

saxenam06 · 2022-04-29T12:33:16Z

Thanks a lot for your feedback. After pressing space bar I start getting the following message. Could you please have a quick glance below and give me a feedback again what could be going wrong. Many Thanks again for your help.

NEW GAME ITERATION__________
Traceback (most recent call last):

File "C:\Users\pulom\Downloads\GPG-drive-master\GPG-drive-master\examples\merging_

scenario_paper.py", line 126, in
vis.run()

File "C:\Users\pulom\Downloads\GPG-drive-master\GPG-drive-master\src\visualize.py", line 725, in run
pyglet.app.run()

File "C:\Users\pulom\anaconda3\lib\site-packages\pyglet\app_init_.py", line 107, in run
event_loop.run()

File "C:\Users\pulom\anaconda3\lib\site-packages\pyglet\app\base.py", line 167, in run
timeout = self.idle()

File "C:\Users\pulom\anaconda3\lib\site-packages\pyglet\app\base.py", line 237, in idle
redraw_all = self.clock.call_scheduled_functions(dt)

File "C:\Users\pulom\anaconda3\lib\site-packages\pyglet\clock.py", line 292, in call_scheduled_functions
item.func(now - item.last_ts, *item.args, **item.kwargs)

File "C:\Users\pulom\Downloads\GPG-drive-master\GPG-drive-master\src\visualize.py", line 304, in control_loop
vehicle.control(steer, gas)

File "C:\Users\pulom\Downloads\GPG-drive-master\GPG-drive-master\src\car.py", line 767, in control
self.initialize_solvers()

File "C:\Users\pulom\Downloads\GPG-drive-master\GPG-drive-master\src\car.py", line 747, in initialize_solvers
self.optimizer = lagrangiansolver.solver(self.id, self.players, self.shared_reward, self.g, self.h, self.v, self.lbv, self.ubv,

File "C:\Users\pulom\Downloads\GPG-drive-master\GPG-drive-master\src\lagrangiansolver.py", line 144, in init
self.solver = src.solvers.get_panocpy_solver(self.problem, self.solver_settings, self.bounds, self.id)

File "C:\Users\pulom\Downloads\GPG-drive-master\GPG-drive-master\src\solvers.py", line 186, in get_panocpy_solver
panocpy_prob = pa.generate_and_compile_casadi_problem_full(f_prob, g1_prob, g2_prob, name=name)

File "C:\Users\pulom\anaconda3\lib\site-packages\panocpy\casadi_problem.py", line 415, in generate_and_compile_casadi_problem_full
subprocess.run(configure_cmd, check=True)

File "C:\Users\pulom\anaconda3\lib\subprocess.py", line 528, in run
raise CalledProcessError(retcode, process.args,

CalledProcessError: Command '['cmake', '-B', 'C:\Users\pulom\AppData\Local\Temp\panocpy\cache\build\build', '-S', 'C:\Users\pulom\AppData\Local\Temp\panocpy\cache\build']' returned non-zero exit status 1.

brechtevens · 2022-05-02T07:04:39Z

That is strange, it seems like the provided 'panocpy' package is unable to build the quadratic penalty solver. Basically, as you probably have noticed, I am using a forked version of PANOC-ALM as my solver since the numerical experiments seem to perform better when using a quadratic penalty method instead of ALM.

Could you try the following things?

use ALM instead of quadratic penalty, by adding the following options to the merging scenario:
- experiment.solver_settings.panoc_use_alm = True
- experiment.learning_settings.panoc_use_alm = True
use OpEn instead of panocpy, by using
- experiment.solver_settings.solver = 'OpEn'
- experiment.learning_settings.solver = 'OpEn'

Finally, some additional questions:

What operating system are you using?
Are you using Python 3.9?

saxenam06 · 2022-05-03T13:14:00Z

Hello, Many Thanks again for your inputs and time. I am working on Windows 11 Home, with Anaconda and using Python 3.9.7 version.
Please also see this notebook to have a quick galnce of my installation procedure.
https://github.com/saxenam06/GPG-drive/blob/master/GPG_Drive_Installed.ipynb
By using the first proposal to use panoc_alm I get the same error as above mentioned.
By using the second proposal of using OpEn solver, I get some progress but then get the following error.
NEW GAME ITERATION
[INFO] Initialising builder
[INFO] Checking user parameters
[INFO] Creating necessary folders
[INFO] Copying icasadi interface to target directory
[INFO] Generating icasadi's Cargo.toml
[INFO] Generating Cargo.toml for target optimizer
[INFO] Generating icasadi Rust library file
Traceback (most recent call last):

File "C:\Users\pulom\Downloads\GPG-drive-master\GPG-drive-master\examples\merging_scenario_paper.py", line 129, in
vis.run()