target.evaluator.default {irace}

target.evaluator.default

Description

target.evaluator.default is the default targetEvaluator function that is invoked if targetEvaluator is a string (by default targetEvaluator is NULL and this function is not invoked). You can use it as an advanced example of how to create your own targetEvaluator function.

Usage

target.evaluator.default(
  experiment,
  num.configurations,
  all.conf.id,
  scenario,
  target.runner.call
)

Arguments

experiment

A list describing the experiment. It contains at least:

id.configuration

An alphanumeric string that uniquely identifies a configuration;

id.instance

An alphanumeric string that uniquely identifies an instance;

seed

Seed for the random number generator to be used for this evaluation, ignore the seed for deterministic algorithms;

instance

String giving the instance to be used for this evaluation;

bound

(only when capping is enabled) Time bound for the execution;

configuration

1-row data frame with a column per parameter name;

switches

Vector of parameter switches (labels) in the order of parameters used in configuration.

num.configurations

Number of configurations alive in the race.

all.conf.id

Vector of configuration IDs of the alive configurations.

scenario

(list())
Data structure containing irace settings. The data structure has to be the one returned by the function defaultScenario() or readScenario().

target.runner.call

String describing the call to targetRunner that corresponds to this call to targetEvaluator. This is used for providing extra information to the user, for example, in case targetEvaluator fails.

Value

The function targetEvaluator must return a list with one element "cost", the numerical value corresponding to the cost measure of the given configuration on the given instance.

The return list may also contain the following optional elements that are used by irace for reporting errors in targetEvaluator:

error

is a string used to report an error;

outputRaw

is a string used to report the raw output of calls to an external program or function;

call

is a string used to report how targetRunner called an external program or function.

Author(s)

Manuel López-Ibáñez and Jérémie Dubois-Lacoste