VAMOS: Vectorized Architecture for Multiobjective Optimization Studies

A high-performance, unified framework for Multi-Objective Evolutionary Algorithms (MOEA) in Python.

Official repository: vamos-optimization/VAMOS. Issues, pull requests, releases, and development are coordinated there. See repository governance.

VAMOS bridges the gap between simple research scripts and large-scale optimization studies. It provides a unified API for running state-of-the-art algorithms across diverse problems, backed by vectorized kernels with NumPy as the exact reference path and optional Numba acceleration for core kernels.

VAMOS 1.0.0 is the first official public release and compatibility baseline. VAMOS 1.1.0 corrects constraint handling and IBEA's hypervolume indicator, adds durable batches and statistical comparison to Studio, and lets explicit engine="auto" prefer Numba across all nine built-in algorithms. See the 1.1.0 release notes, including the impact on constrained and hypervolume-based IBEA experiments, and the earlier 1.0.2 release notes. Earlier version strings and Git tags were internal pre-public development markers, not prior public releases. See the stability policy and known limitations.

Key Features

• Unified API: A clear, fluent interface vamos.optimize() for all workflows.
• Battle-Tested Algorithms: NSGA-II/III, MOEA/D, SMS-EMOA, SPEA2, IBEA, SMPSO, AGE-MOEA, RVEA.
• Unified Archiving: Consistent external archive configuration via .external_archive(capacity=..., pruning=...), with bounded or unbounded archives and pruning policies crowding, hv, mc_hv, knn, maxmin, and ref_dirs. When an external archive is enabled, top-level results come from it by default unless result_mode="population" is requested.
• Multi-Fidelity Tuning: Hyperband-style racing with warm-start checkpoints for sample-efficient algorithm configuration.
• Ready-to-use Tuning Backends: racing and random work out of the box; install the optional tuning extra to enable optuna, bohb_optuna, smac3, and bohb via vamos tune.
• Performance Driven: Vectorized NumPy kernels with optional Numba JIT acceleration for core kernels.
• Interactive Analysis: Built-in dashboards with explore_result_front(result) and publication-ready LaTeX tables.
• Visual Problem Builder: Define custom problems in the experimental VAMOS Studio; local Python preview requires an explicit trusted-code opt-in.
• Extensible: Standardized protocols for adding custom problems, operators, and algorithms.

Canonical customization guide: docs/topics/extending.md.

Quick Install

pip install vamos-optimization

For development and extras (Windows PowerShell):

# Create virtual environment
python -m venv .venv
.\.venv\Scripts\Activate.ps1

# Install core + essential extras
pip install "vamos-optimization[compute,research,analysis]"

For development and extras (Linux / macOS):

# Create virtual environment
python -m venv .venv
source .venv/bin/activate

# Install core + essential extras
pip install "vamos-optimization[compute,research,analysis]"

Optional model-based tuning backends (optuna, smac3, bohb):

pip install "vamos-optimization[tuning]"

smac3 in VAMOS is provided by the PyPI package smac (SMAC3):

pip install "smac>=2.0"

Backend Capability Matrix

In VAMOS 1.1.0, use engine="auto" (CLI: --engine auto) to prefer loadable Numba for all nine built-in algorithms, with a NumPy fallback. Numba remains optional. Omitting engine still selects NumPy, preserving the 1.x default; explicit engine="numba" fails if unavailable. First use can spend more time compiling than a short optimization takes. See the selection and replay policy before comparing timings or moving runs between environments.

Operator shorthand accepts either numeric probabilities or the string literal "1/n" for per-variable mutation rates.

API Contracts

• max_evaluations is a strict evaluation budget. It must be at least the resolved population size, and the final generation is truncated when needed so budget-terminated runs report exactly result.data["evaluations"] == max_evaluations.
• Public algorithm results expose evaluations as the evaluation-count metric. Internal checkpoints may still store n_eval.
• Hypervolume utilities require the reference point to dominate all points by default. Pass allow_ref_expand=True only when automatic reference-point expansion is intended.
• NSGA-III, RVEA, and MOEA/D require pop_size to match their configured reference directions or weight lattice; incompatible configurations fail early with an actionable error.

Quickstart

Solve the ZDT1 benchmark problem with NSGA-II in just a few lines:

from vamos import optimize

result = optimize(
    "zdt1",
    algorithm="nsgaii",
    max_evaluations=10000,
    pop_size=100,
    engine="numpy",
    seed=42,
)

front = result.front()
print(f"Non-dominated solutions: {len(front) if front is not None else 0}")

Prefer a guided CLI? Run:

vamos quickstart

This wizard writes a reusable config and stores results under results/quickstart/.

Use vamos quickstart --template list to inspect domain templates.

New to Python? Start with the Minimal Python Track: docs/guide/minimal-python.md.

After a run, summarize results with:

vamos summarize --results results/quickstart

Inspect, fully verify, or exactly replay a canonical built-in run:

vamos results inspect RUN_DIR
vamos results verify RUN_DIR --require-level exact
vamos reproduce RUN_DIR

For a small study in one call:

study = optimize("zdt1", algorithm="nsgaii", max_evaluations=4000, seed=[0, 1, 2])
print(study.mean("evaluations"))
print(study.best_run("evaluations").meta["seed"])

All functionality lives under one command. Run vamos help to list everything:

Tuning Quick Start

You can use the implemented tuning backends directly from vamos tune: racing, random, optuna, bohb_optuna, smac3, bohb.

Check backend availability in your current environment:

vamos tune --list-backends

Quick verification with the built-in backend and tiny budgets:

vamos tune --instances zdt1,zdt2,zdt3,dtlz1,dtlz2,wfg1 --algorithm nsgaii --backend random --smoke --output-dir report/tuning_smoke

Note: racing and random require no extra dependencies. The model-based backends (optuna, bohb_optuna, smac3, bohb) require the optional tuning extra: pip install "vamos-optimization[tuning]". The smac3 backend uses the smac package.

For paper-grade comparisons against external frameworks, do not rely on the open-ended research extra alone. Record a pinned environment or lockfile alongside reported results.

Recommended robust command (fallback + suite-stratified split):

vamos tune \
  --instances zdt1,zdt2,zdt3,dtlz1,dtlz2,wfg1 \
  --algorithm nsgaii \
  --backend optuna \
  --backend-fallback random \
  --split-strategy suite_stratified \
  --budget 5000 \
  --tune-budget 200 \
  --n-jobs -1

Canonical tuning reference: docs/topics/tuning.md.

New to hands-on learning? Open the interactive tutorial notebook:

jupyter notebook notebooks/0_basic/05_interactive_tutorial.ipynb

Define Your Own Problem

Use make_problem() to turn any Python function into a VAMOS-compatible problem -- no classes, no protocols, no NumPy vectorization required:

from vamos import make_problem, optimize

problem = make_problem(
    lambda x: [x[0], (1 + x[1]) * (1 - x[0] ** 0.5)],
    n_var=2,
    n_obj=2,
    bounds=[(0, 1), (0, 1)],
    encoding="real",
)

result = optimize(problem, algorithm="nsgaii", max_evaluations=5000, seed=42)

Your function receives a single solution x (array of length n_var) and returns a list of n_obj objective values. When vectorized=False, VAMOS adapts that scalar callable by evaluating one row at a time.

For actual batch performance, pass vectorized=True and write a function that handles (N, n_var) batches directly.

Prefer a file template? The CLI wizard scaffolds a ready-to-run .py file:

vamos create-problem
# Prompts for: name, variables, objectives, bounds, style
# Generates a .py file with TODO markers -- fill in your math and run it

Or use the experimental visual builder in VAMOS Studio -- review your objectives in the browser, pick an algorithm, and see the Pareto front update on each run:

vamos studio
# Open the "Problem Builder" tab

See docs/dev/add_problem.md for all approaches (function, class, or registry).

VAMOS Assist (no-code workflow)

VAMOS Assist provides an end-to-end no-code flow for creating validated experiment plans, materializing runnable projects, and optionally running smoke checks. You can start with deterministic templates (no API keys), and optionally use provider-backed auto planning.

vamos assist go "template-first example" --template demo --smoke

pip install vamos-optimization[openai]
setx OPENAI_API_KEY "..."
vamos assist go "..." --mode auto --provider openai --smoke

See docs/assist.md for the full guide (billing, privacy, artifacts, troubleshooting).

Preferred path: start with optimize(...). Use config objects only when you need fully specified, reproducible runs or plugin algorithms. See docs/guide/getting-started.md for a short decision guide.

Advanced path (explicit config objects):

from vamos import optimize
from vamos.algorithms import NSGAIIConfig
from vamos.problems import ZDT1

problem = ZDT1(n_var=30)
algo = NSGAIIConfig.default(pop_size=100, n_var=problem.n_var)

result = optimize(
    problem,
    algorithm="nsgaii",
    algorithm_config=algo,
    max_evaluations=10000,
    seed=42,
    engine="numpy",
)

Reminder: plain dict configs are intentionally not accepted (use GenericAlgorithmConfig for plugin algorithms).

For comparative evidence against pymoo on fixed seeded cases, run:

python tools/benchmark_compare_pymoo.py --output artifacts/performance/pymoo_comparison.json --markdown artifacts/performance/pymoo_comparison.md

Notes

• For reproducible results, set seed; NumPy/Numba/MooCore backends share the same RNG-driven stochastic operators.
• Troubleshooting guide: docs/guide/troubleshooting.md.
• Algorithm-specific notes (reference directions, operator defaults): docs/reference/algorithms.md.
• Release packaging smoke checklist: docs/release_smoke.md.

Examples & Notebooks

VAMOS comes with a comprehensive suite of Jupyter notebooks organized by tier:

• 0. Basic: Essential concepts and API basics.
  • notebooks/INDEX.ipynb -- Maintained catalog of the full learning surface
  • notebooks/0_basic/01_quickstart.ipynb -- First optimization run
  • notebooks/0_basic/05_interactive_tutorial.ipynb -- Guided hands-on walkthrough
  • notebooks/0_basic/06_optuna_tuning_basics.ipynb -- Introductory Optuna tuning
• 1. Intermediate: Real-world problems, constraints, and deeper analysis.
  • notebooks/1_intermediate/10_discrete_problems.ipynb -- Binary, integer, and permutation encodings
  • notebooks/1_intermediate/11_constrained_optimization.ipynb -- Constraint handling
  • notebooks/1_intermediate/15_mcdm.ipynb -- Multi-criteria decision making
  • notebooks/1_intermediate/16_interactive_explorer.ipynb -- Interactive Pareto front explorer
  • notebooks/1_intermediate/19_algorithm_families_beyond_nsgaii.ipynb -- Current recipes for the other VAMOS MOEAs
• 2. Advanced: Custom extensions, tuning, and research benchmarks.
  • notebooks/2_advanced/21_programmatic_tuning.ipynb -- Canonical Optuna tuning workflow
  • notebooks/2_advanced/23_backends_and_performance.ipynb -- Backend tradeoffs and benchmarking
  • notebooks/2_advanced/30_paper_benchmarking.ipynb -- Inspect retained benchmark data
  • notebooks/2_advanced/27_operator_efficacy.ipynb -- Operator efficacy analysis
  • notebooks/2_advanced/32_ablation_planning.ipynb -- Ablation studies
  • notebooks/2_advanced/33_optuna_tuning_advanced.ipynb -- Multi-fidelity and persistent Optuna workflows
  • notebooks/2_advanced/34_extension_workflows.ipynb -- Custom problem and custom algorithm extension patterns

Tooling Ecosystem

All tools are available as vamos <subcommand>. Run vamos help for the full list.

• vamos profile: Analyze the performance overhead of your experiments.

  vamos profile --problem zdt1 --engines numpy,numba --budget 2000 --output report/profile.csv
  

• vamos bench: Generate full reports comparing multiple algorithms, plus jMetalPy-compatible lab outputs (summary/lab/QualityIndicatorSummary.csv, Wilcoxon tables, boxplots). Boxplots require matplotlib.

  vamos bench ZDT_small --algorithms nsgaii moead --output report/
  

  vamos bench ZDT_small --algorithms nsgaii --output report/ --smoke
  

• vamos tune: You can use the implemented tuners directly from CLI (racing, random, optuna, bohb_optuna, smac3, bohb). --tune-budget counts configuration evaluations; --budget is per-run evaluations.

  vamos tune --problem zdt1 --algorithm nsgaii --budget 5000 --tune-budget 200 --n-seeds 5
  

  • Quick verification path:

    vamos tune --instances zdt1,zdt2,zdt3,dtlz1,dtlz2,wfg1 --algorithm nsgaii --backend random --smoke --output-dir report/tuning_smoke
    

  • Recommended robust invocation (backend fallback + suite-stratified split):

    vamos tune --instances zdt1,zdt2,zdt3,dtlz1,dtlz2,wfg1 --algorithm nsgaii --backend optuna --backend-fallback random --split-strategy suite_stratified --budget 5000 --tune-budget 200 --n-jobs -1
    

• Full tuning reference (canonical docs): docs/topics/tuning.md.
• vamos check: Verify your installation and backend availability.

Citation

The retained benchmark data guide documents the available scientific inputs and their provenance. Manuscript sources and author revision materials are maintained separately from this public repository.

If you use VAMOS in published work, cite it directly:

@software{vamos_2026,
  title = {VAMOS: Vectorized Architecture for Multiobjective Optimization Studies},
  author = {Rodriguez Uribe, Nicolas and Herr{\'a}n, Alberto and Nebro, Antonio J. and Del Ser, Javier and Colmenar, J. Manuel},
  year = {2026},
  version = {1.1.0},
  url = {https://github.com/vamos-optimization/VAMOS}
}

The maintained citation metadata lives in CITATION.cff.

Contributing

We welcome contributions! Please see CONTRIBUTING.md for guidelines.

• Found a bug? Open an issue.
• Want to add an algorithm? Check dev/add_algorithm.md in the docs.
• Using AI tools? Read .agent/docs/AGENTS.md for our AI coding standards.
• Troubleshooting: docs/guide/troubleshooting.md.
• Security issues: See SECURITY.md for private reporting.
• Contributors: See AUTHORS.md.

VAMOS is a research-oriented multi-objective optimization framework.