Package org.uma.jmetal.util.observer.impl

Class FrontPlotObserver<S extends Solution<?>>

java.lang.Object

org.uma.jmetal.util.observer.impl.FrontPlotObserver<S>

Type Parameters:

S - Type of the solutions being observed, must extend Solution<?>

All Implemented Interfaces:

Observer<Map<String,Object>>

public class FrontPlotObserver<S extends Solution<?>>
extends Object
implements Observer<Map<String,Object>>

An observer that displays the current Pareto front during the execution of a multi-objective optimization algorithm. It visualizes the solutions in a 2D scatter plot, showing how the front evolves over time.

This observer is typically used to monitor the progress of multi-objective optimization algorithms by plotting the objective values of the current population. It can also display reference fronts for comparison and filter out dominated solutions if needed.

Example usage:

 FrontPlotObserver<Solution<?>> observer = new FrontPlotObserver<>(
     "Front Plot", "Objective 1", "Objective 2", "Population", 100);
 algorithm.getObservable().register(observer);
 

Author:

Antonio J. Nebro

Constructor Summary

Constructors

Constructor	Description
FrontPlotObserver(String title, String xAxisTitle, String yAxisTitle, String legend, int plotUpdateFrequency)	Creates a new FrontPlotObserver instance.

Method Summary

Modifier and Type	Method	Description
org.knowm.xchart.XYChart	chart()	Returns the underlying XYChart instance for advanced customization.
void	filterDominatedSolutions(boolean filterDominatedSolutions)	Enables or disables filtering of dominated solutions.
void	setFront(double[][] front, String frontName)	Adds a reference front to the plot for comparison with the current population.
void	setPoint(double x, double y, String pointName)	Adds a single point to the plot, which can be used to mark specific solutions or reference points.
void	update(Observable<Map<String,Object>> observable, Map<String,Object> data)	Updates the plot with the current population from the optimization algorithm.

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Details

FrontPlotObserver

public FrontPlotObserver(String title,
 String xAxisTitle,
 String yAxisTitle,
 String legend,
 int plotUpdateFrequency)

Creates a new FrontPlotObserver instance.

Parameters:

title - The title of the plot

xAxisTitle - Title for the X-axis (typically first objective name)

yAxisTitle - Title for the Y-axis (typically second objective name)

legend - Legend text for the population points

plotUpdateFrequency - How often to update the plot (in number of evaluations)

Throws:

IllegalArgumentException - if plotUpdateFrequency is less than 1

Method Details

setFront

public void setFront(double[][] front,
 String frontName)

Adds a reference front to the plot for comparison with the current population. The reference front is typically a known Pareto-optimal front or another reference set of solutions.

Parameters:

front - 2D array where each row represents a solution in the reference front

frontName - The name to display in the legend for this front

Throws:

IllegalArgumentException - if front is null or empty

setPoint

public void setPoint(double x,
 double y,
 String pointName)

Adds a single point to the plot, which can be used to mark specific solutions or reference points.

Parameters:

x - X-coordinate of the point (first objective value)

y - Y-coordinate of the point (second objective value)

pointName - The name to display for this point in the legend

filterDominatedSolutions

public void filterDominatedSolutions(boolean filterDominatedSolutions)

Enables or disables filtering of dominated solutions. When enabled, only non-dominated solutions from the current population will be displayed.

Parameters:

filterDominatedSolutions - If true, only non-dominated solutions will be plotted

update

public void update(Observable<Map<String,Object>> observable,
 Map<String,Object> data)

Updates the plot with the current population from the optimization algorithm. This method is called automatically by the observed algorithm.

Specified by:

update in interface Observer<S extends Solution<?>>

Parameters:

observable - The observable object (typically the optimization algorithm)

data - Map containing the current state with the following expected keys: - "EVALUATIONS": Current number of evaluations (Integer) - "POPULATION": Current population of solutions (List)

Throws:

ClassCastException - if the data map contains values of unexpected types

chart

public org.knowm.xchart.XYChart chart()

Returns the underlying XYChart instance for advanced customization. Use this method to modify chart properties not exposed by the FrontScatterPlot API.

Returns:

The XYChart instance used for plotting