NESO is a local-search solver for nonlinear optimization. This release and documentation are temporarily limited to PUBO. Broader problem support will be documented when added.
NESO <model.gms> [max_iters] [seed] [solution_file]
| Argument | Description |
|---|---|
model.gms |
Path to a compatible .gms file (required). |
max_iters |
Maximum number of local-search iterations. Default: 100000. |
seed |
RNG seed. Use 0 for a nondeterministic run (entropy-based); any nonzero value makes the run reproducible for that seed. |
solution_file |
Optional. If provided, writes the best solution found to this path after the run. |
Prints one block per run, including:
file:— input pathn_vars:/n_terms:— parsed problem sizestart_obj(all-zero):— objective at the all-zero startbest_obj:— best objective found
./NESO /path/to/instance.gms./NESO /path/to/instance.gms 10000 42./NESO /path/to/instance.gms 10000 12345 /path/to/output.sol./NESO /path/to/instance.gms 10000 0 /path/to/output.solMake the binary executable if needed: chmod +x NESO.