Skip to content

VAMOS · Scientific Python

Multi-objective optimizationbuilt for reproducible studies.

Compare algorithms, solve your own problems, and preserve the evidence behind every run.

Get started API reference
$python -m pip install "vamos-optimization==1.0.0" 
from vamos import optimize

result = optimize(
    "zdt1",
    algorithm="nsgaii",
    pop_size=40,
    max_evaluations=400,
    engine="numpy",
    seed=42,
)

front = result.front()
40 solutions · 2 objectives
Actual NSGA-II run on ZDT1: grey circles show the final population, and teal crosses identify its non-dominated subset.
Real output from the code shown: 400 evaluations, seed 42. A first run, not a convergence benchmark. Read and reproduce this result →
01

Try VAMOS

Run a built-in problem and inspect its trade-offs.

Quickstart → 02

Solve my problem

Define your decisions, objectives, bounds, and constraints.

Custom problems → 03

Run a reproducible study

Plan a study, check its budget, and trace every result.

Reproducible studies →

Prefer runnable source? Run the three executable journeys.

From a first run to a scientific study

VAMOS 1.0 distinguishes stable optimization, run-artifact, and single-owner study surfaces from experimental features such as Studio, provider integrations, and tuning. Check Stability and versioning before depending on an API as a 1.x compatibility commitment.

AlgorithmsNSGA-II/III, MOEA/D, SMS-EMOA, SPEA2, IBEA, SMPSO, AGE-MOEA, RVEA.
EncodingsReal, integer, binary, permutation, and mixed decision variables where supported.
BackendsNumPy reference execution with optional Numba kernel and MooCore indicator acceleration.
Research toolingStudies, benchmarking, tuning, result inspection, verification, replay, and analysis.

For precise signatures and configuration contracts, use the API reference, algorithm reference, and problem reference.

Project and citation

Citation metadata is maintained in CITATION.cff. Security policy and supported reporting channels are maintained in SECURITY.md. See Known limitations, the roadmap, and repository governance for project-level information.