Weird Output

[ParthavK]

Hi,
I am testing the solver using:
Input file: input_25.json
Generated output: solution_25.json

Command used:

./optimization/packingsolver_irregular \
  --verbosity-level 1 \
  --input solver_io/input_25.json \
  --time-limit 1000 \
  --certificate solver_io/solution_25.json \
  --optimization-mode Anytime

The solver runs successfully and generates solution_25.json.

However, I want to confirm whether:

• The objective function in input_25.json is correctly defined?
• The certificate output (solution_25.json) is being interpreted correctly?
• Is there anything wrong in my configuration (input format / objective / parameters)?
• Could someone please confirm whether this setup is correct or if I am missing something?

Thank you.

[fontanf]

Hi,

I think you're not using the latest version of the solver.
There is a bug with the latest version that makes the solver throw an exception with the VariableSizedBinPacking objective.
With the latest version of the solver with the bug fixed, I get this solution:

I'll push the fix soon.

Considering the input, are you sure that you want to use the VariableSizedBinPacking objective? This objective only optimizes the cost of the bins used. It doesn't optimize the leftover.
Here is the solution I get with the BinPackingWithLeftovers objective:

[fontanf]

Also note that in the latest version, item_bin_minimum_spacing must be defined in bin_types for each bin type

[ParthavK]

Hi,
Thank you for helping me out.
I am using variable-sized bin packing because I have multiple bins with different sizes. That’s why I chose this configuration.
My main goal is to optimize the bin packing with leftover as efficiently as possible. If there is a better modeling approach, parameter configuration, or optimization strategy that would give better results, I would really appreciate your suggestion.
Thanks again for your guidance.

[fontanf]

Then you can try with two successive solves: the first one with the variable-sized bin packing objective to get the cheapest bin combination; and a second one with the bin packing with leftovers objective and only the bins selected in the previous step to get a more compact solution

[ParthavK]

Thank you for your help. Yes, that’s exactly what I was thinking as well.