desdeo.optimization

This package contains methods for solving single-objective optimisation problems. These are contained in OptimizationMethod. It also contains scalarisation functions, used for converting multi-objective problems into single-objective fucntions. These are contained in OptimizationProblem. Both are used as primitives by the methods defined in desdeo.method.

class desdeo.optimization.PointSearch(optimization_problem)

Bases: desdeo.optimization.OptimizationMethod.OptimizationMethod

__abstractmethods__ = frozenset({})

__module__ = 'desdeo.optimization.OptimizationMethod'

_search(**params)

The actual search for the optimal solution

This is an abstract class that must be implemented by the subclasses

Parameters

**params (dict [optional]) – Parameters for single objective optimization method

Return type

Tuple[ndarray, List[float]]

class desdeo.optimization.SciPy(optimization_problem)

Bases: desdeo.optimization.OptimizationMethod.OptimalSearch

Optimal search using scipy.optimize.minimize().

__abstractmethods__ = frozenset({})

__module__ = 'desdeo.optimization.OptimizationMethod'

_const(x, *ncon)

_objective(x)

Return objective function value

Parameters

x (list of values) – Decision variable vector to be calclated

_search(max=False, **params)

The actual search for the optimal solution

This is an abstract class that must be implemented by the subclasses

Parameters

**params (dict [optional]) – Parameters for single objective optimization method

Return type

Tuple[ndarray, List[float]]

class desdeo.optimization.SciPyDE(optimization_problem)

Bases: desdeo.optimization.OptimizationMethod.OptimalSearch

Optimal search using scipy.optimize.differential_evolution().

__abstractmethods__ = frozenset({})

__init__(optimization_problem)

Initialize self. See help(type(self)) for accurate signature.

__module__ = 'desdeo.optimization.OptimizationMethod'

_objective(x)

Return objective function value

Parameters

x (list of values) – Decision variable vector to be calclated

_search(**params)

The actual search for the optimal solution

This is an abstract class that must be implemented by the subclasses

Parameters

**params (dict [optional]) – Parameters for single objective optimization method

Return type

Tuple[ndarray, List[float]]