Optimization (CADETProcess.optimization)

Optimization (CADETProcess.optimization)#

The optimization module provides functionality for minimizing (or maximizing) objective functions, possibly subject to constraints. It includes interfaces to several optimization suites, notably, scipy.optimize and pymoo.

OptimizationProblem#

OptimizationProblem(name)

Class for configuring optimization problems.

Optimizer#

Base#

OptimizerBase(progress_frequency, x_tol, ...)

BaseClass for optimization solver APIs.

Scipy#

TrustConstr(gtol, xtol, barrier_tol, ...)

Wrapper for the trust-constr optimization method from the scipy optimization suite.

COBYLA(rhobeg, tol, maxiter, disp, catol, ...)

Wrapper for the COBYLA optimization method from the scipy optimization suite.

COBYQA(disp, maxfev, maxiter, f_target, ...)

Wrapper for the COBYQA optimization method from the scipy optimization suite.

NelderMead(maxiter, xatol, fatol, adaptive, ...)

Wrapper for the Nelder-Mead optimization method from the scipy optimization suite.

SLSQP(ftol, eps, disp, maxiter, iprint, ...)

Wrapper for the SLSQP optimization method from the scipy optimization suite.

Pymoo#

NSGA2(seed, pop_size, xtol, ftol, cvtol, ...)

NSGA2 Algorithm.

U_NSGA3(seed, pop_size, xtol, ftol, cvtol, ...)

U-NSGA3 Algorithm.

Ax#

BotorchModular(acquisition_fn, ...)

Modular bayesian optimization algorithm.

GPEI(early_stopping_improvement_window, ...)

Gaussian Process with Expected Improvement for single objectives.

NEHVI(early_stopping_improvement_window, ...)

Noisy expected hypervolume improvement multi-objective algorithm.

Population#

Individual(x, x_transformed, cv_bounds, ...)

Set of variables evaluated during Optimization.

Population([id])

Collection of Individuals evaluated during Optimization.

Results#

OptimizationResults(success, exit_flag, ...)

Optimization results.

Cache#

ResultsCache([use_diskcache, directory, ...])

Cache to store (intermediate) results.

ParallelizationBackend#

ParallelizationBackendBase(n_cores)

Base class for all parallelization backend adapters.

SequentialBackend(n_cores)

Sequential execution backend for evaluating the target function.

Joblib(verbose, n_cores)

Parallelization backend implementation using joblib.

Pathos(n_cores)

Parallelization backend using the pathos library.