Keyboard shortcuts

Press ← or → to navigate between chapters

Press S or / to search in the book

Press ? to show this help

Press Esc to hide this help

fcmaes-rust documentation

This directory documents the standalone Rust implementation. Generated API documentation comes from the Rust sources; these guides focus on architecture, configuration, workflows, and runnable examples.

Read the guides as a navigable mdBook site, or browse the same canonical Markdown files in the repository. Exact public signatures, runnable API examples, and primary algorithm references are in the fcmaes-core API reference.

Documentation map

DocumentRead this for
AI problem-solving contextSelecting algorithms, parameters, budgets, encodings, and validation for a new user problem
FoundationsSeven compact lessons, standard academic suites, Lennard-Jones scaling with a required gradient reference, analytic fronts, and audited indicators
Getting startedBuilding, testing, generating rustdoc, and running a first optimizer
Choosing an optimizerDeciding whether fcmaes-rust, a structured solver, gradients, a surrogate, or another search representation fits the problem
The optimizer boundaryWhy local, Bayesian, gradient, and structured solvers remain external, with corrected DE/NM/BO experiments
ArchitectureWorkspace layout, execution paths, concurrency, and scope
OptimizersPure-Rust optimizer APIs, defaults, one-shot operation, and ask/tell operation
RetryBasic, coordinated, and multi-objective retry
Optional Python bindingsDirect PyO3 extension surface and GIL considerations
ExamplesEvery native binary, data input, GTOP problem, monitor, and benchmark
Combinatorial encodingsTurning fixed real vectors into integers, categories, subsets, permutations, partitions, and repaired schedules
Buckingham–PiNumerical dimensional analysis, holdout scoring, BiteOpt retry, and MODE
Application tutorialsTwenty-two native optimization applications, including simulation, astrodynamics, circuit analysis, structural topology, policy search, ML hyperparameter tuning, embedded LPs, and custom-backend verification
DevelopmentFormatting, linting, tests, coverage, rustdoc, and extension points

Implemented Rust surface

  • Bounded fitness handling, normalization, scalar and population evaluation, evaluation counting, and PCG-based random generation.
  • Differential Evolution, active CMA-ES, CR-FM-NES, PGPE, Dual Annealing, BiteOpt, MODE, CVT-MAP-Elites, and the Diversifier.
  • Independent retry, coordinated advanced retry, and weighted multi-objective retry.
  • Native GTOP and Mazda objectives plus application drivers for factory design, stock trading, material flow, flexible job-shop/harvesting, multi-UAV task assignment, Buckingham–Pi analysis, spherical t-design, transfer scheduling, damped control, F-8, and Lotka-Volterra.
  • Tested real-vector decoders for integers, categories, Booleans, random-key permutations, fixed-cardinality subsets, partitions, and ordered times.
  • Twenty-two standalone native application tutorials, including a PGPE/CR-FM-NES neural policy-search showcase, staged pykep-core GTOC1 optimization, validated thevenin transient gate-driver optimization, validation-aware SmartCore hyperparameter optimization, and robust room-ventilation optimization with a purpose-built D2Q9/D2Q5 backend, held-out releases, MODE, and MAP-Elites.
  • A standalone foundations workspace with classic, ZDT, DTLZ, and Lennard-Jones suites, explicit CEC/WFG/BBOB evidence gates, and exact or typed Monte Carlo quality indicators.
  • An optional PyO3 extension distributed through the fcmaes_rust Python facade.

Fast path

From the repository root:

cargo test --workspace
cargo build --release --workspace
cargo doc --workspace --no-deps

To build the guide and tutorial site locally after installing mdBook 0.5.4:

python scripts/build_book.py
python scripts/check_book_links.py

The assembled source and output are written below target/.

Run a small native GTOP workload:

cargo run --release -p fcmaes-examples --bin gtop-examples -- \
  --problem cassini1 --retries 16 --evaluations 5000 --workers 16 --seed 1

Run the hard Messenger Full workload with live progress:

cargo run --release -p fcmaes-examples --bin gtop-advexamples -- \
  --problem messenger-full --retries 50000 --evaluations 1500 \
  --workers 16 --seed 1 --value-limit 12 \
  --max-eval-fac 50 --check-interval 100 --progress-interval 10