Package org.moeaframework.analysis.tools


package org.moeaframework.analysis.tools
Command line tools for evaluating MOEAs and analyzing results. Most of these tools are designed for producing control maps and performing sensitivity analysis on parameters. These are largely based on the paper cited below.

References:

  1. Hadka, D. and Reed, P. "Diagnostic Assessment of Search Controls and Failure Modes in Many-Objective Evolutionary Optimization." Evolutionary Computation, 20(3):423-452, 2012.
  • Classes
    Class
    Description
    Command line utility for calculating the best, probability of attainment, efficiency and controllability metrics.
    Converts a result file into an ARFF file that can be loaded into various data mining software (e.g., Weka).
    Command line utility for evaluating an algorithm using many parameterizations.
    Command line utility for extracting data from a result file.
    Command line utility for negating objective values in result files.
    Utility for merging two or more populations identified by unique sources and determining the contribution of each source to the combined non-dominated population.
    Command line utility for evaluating the approximation sets stored in a result file and computing its metric file.
    Command line utility for counting the number of entries in a result file.
    Command line utility for merging the approximation sets stored in one or more result files.
    Command line utility for merging the approximation sets in one or more result files across the seeds.
    Command line utility for evaluating an algorithm using many parameterizations.
    Command line utility for producing randomly-generated parameters for use by the Evaluator or RuntimeEvaluator.
    Command line utility for reporting the number of solutions in a set that are contained within a reference set.
    Command line utility for generating reference sets for a given problem.
    Command line utility for calculating the hypervolume of approximation sets.
    Command line utility for computing statistics across multiple data files.
    Global sensitivity analysis of blackbox model output using Saltelli's improved Sobol' global variance decomposition procedure.
    Command line utility for solving an optimization problem using any of the supported optimization algorithms.