org.drools.planner.core.score.definition
Interface ScoreDefinition<S extends Score>

All Known Implementing Classes:

AbstractScoreDefinition, HardAndSoftScoreDefinition, SimpleDoubleScoreDefinition, SimpleScoreDefinition

public interface ScoreDefinition<S extends Score>

A ScoreDefinition knows how to compare scores and what the perfect maximum/minimum Score is.

See Also:

AbstractScoreDefinition, HardAndSoftScoreDefinition

Method Summary

ScoreCalculator	buildScoreCalculator() TODO remove when the rule that sums the final score can be written as a single rule and ScoreCalculator is dead
double	calculateTimeGradient(S startScore, S endScore, S score) See explanation in Termination.calculateSolverTimeGradient(DefaultSolverScope).
S	getPerfectMaximumScore() The perfect maximum score is the score of which there is no better in any problem instance.
S	getPerfectMinimumScore() The perfect minimum score is the score of which there is no worser in any problem instance.
Score	parseScore(String scoreString) Parses the String and returns a Score.
Double	translateScoreToGraphValue(S score)

Method Detail

getPerfectMaximumScore

S getPerfectMaximumScore()

The perfect maximum score is the score of which there is no better in any problem instance. This doesn't mean that the current problem instance, or any problem instance for that matter, could ever attain that score.

For example, most cases have a perfect maximum score of zero, as most use cases only have negative scores.

Returns:

null if not supported

getPerfectMinimumScore

S getPerfectMinimumScore()

The perfect minimum score is the score of which there is no worser in any problem instance. This doesn't mean that the current problem instance, or any problem instance for that matter, could ever attain such a bad score.

For example, most cases have a perfect minimum score of negative infinity.

Returns:

null if not supported

parseScore

Score parseScore(String scoreString)

Parses the String and returns a Score.

Parameters:

scoreString - never null

Returns:

never null

calculateTimeGradient

double calculateTimeGradient(S startScore,
                             S endScore,
                             S score)

See explanation in Termination.calculateSolverTimeGradient(DefaultSolverScope).

Parameters:

startScore - never null

endScore - never null

score - never null

Returns:

between 0.0 and 1.0

translateScoreToGraphValue

Double translateScoreToGraphValue(S score)

Parameters:

score - never null

Returns:

null if should not be shown on the graph

buildScoreCalculator

ScoreCalculator buildScoreCalculator()

TODO remove when the rule that sums the final score can be written as a single rule and ScoreCalculator is dead

Returns:

never null