Test Evaluators

This section documents the benchmark test evaluators available in the standard_evaluator.evaluators.test module. These evaluators implement well-known mathematical optimization problems with known solutions, useful for testing optimization algorithms or as stand-in functions while developing standard_evaluator workflows.

Each test evaluator has a known_solution property returning the optimal input values and corresponding response values as a DataFrame.

Base Class

class TestEvaluator

Bases: Evaluator

Abstract test function

This defines the abstract test function

Methods:

__init__([name, comp_cost])

Initialize the problem and set the inputs and outputs

Attributes:

`test_info`	Provides a dictionary of information on the test evaluator.
`known_solution`	The known solution of the problem.

__init__(name=None, comp_cost=100, **kwargs)

Initialize the problem and set the inputs and outputs

Parameters:

name (str, optional) – Name to give the evaluator. Defaults to evaluator type name. If a value is provided, this will be appended to the default.
comp_cost (float) – Computational cost of evaluating with this evaluator. Defaults to 100.
kwargs (dict) – Arguments sent to _create_opt_problem()

property test_info: dict

Provides a dictionary of information on the test evaluator.

Returns:: A dictionary containing data describing the evaluator.
Return type:: dict

property known_solution: DataFrame

The known solution of the problem. Provides the site(s) and output values for the problem.

Returns:: DataFrame providing the optimal solution of the problem
Return type:: pd.DataFrame

Categories

Test evaluators are organized by their optimization goal: