More time does not mean better results?

[scorpioares]

Description
I am using LeapHybridCQMSolver to solve the CQM model. If I set the time_limit to None (which is 5s according to the details of problem status), the solver returns satisfying feasible solutions. However, if I set the time_limit to a larger constant such as 30s, the solver cannot find a feasible solution.

To Reproduce
I am unsure whether my model causes this kind of result or something from the SDK and solver. Since my model is under business work, I cannot provide it here.

Expected behavior
I expected a better result by making the time_limit parameter larger, or at least the same result.

Environment:

• OS: [macOS 12.4]
• Python version: [3.10]

Additional context
I wonder if it is a rare case and mostly comes from the model's structure or if it comes from the internal issue of the hybrid solver.

[arcondello]

Hi @scorpioares, thanks for the report! Would it be possible to share the CQM that you submitting? You can save it by running

import shutil

with cqm.to_file() as src_f:
    with open('example.cqm', 'wb') as dest_f:
        shutil.copyfileobj(src_f, dest_f)

[scorpioares]

@arcondello, thank you for your reply. I have saved the CQM file. How can I send it to you? Is there a safer way to do that to protect sensitive business information?