DTLZ problems

This package contains many convenient classes to solve DTLZ problems within ParadisEO-MOEO. Examples using NSGAII, IBEA and SPEA2 are provided.


Download

To install the package, please refer to the "README.txt" file.


Problem description

DTLZ Test Problems

Reference: K. Deb, L. Thiele, M. Laumanns, E. Zitzler: Scalable test problems for evolutionary multi-objective optimization. Evolutionary Multiobjective Optimization: Theoretical Advances and Applications (2005) 105-145.


Solver description

Theses problems are tested by algorithms implemented using ParadisEO-MOEO.

The following components are representation-dependent (some are already defined in ParadisEO):

  • eoRealVectorBounds: Solution encoding
  • eoRealInitBounded: The initializer
  • DTLZXEval (X is the number of the problem 1, 2, 3, 4, 6, 7): Evaluator containing the objective function
  • SBXCrossover: crossover
  • PolynomialMutation: mutation

The others are generic in ParadisEO:

  • eoPop: the population
  • eoGenContinue: Stopping criteria (fixed number of generations)
  • eoTimeContinue: Another stopping criteria (fixed an evaluation time)

The three algorithms used:

  • moeoIBEA
  • moeoNSGAII
  • moeoSPEA2