EDA++

Estimation of Distribution Algorithms with Feasibility Conserving Mechanisms for Constrained Continuous Optimization

Suplementary Materials

This repository contains the codes for the algorithms and the obtained data for the research on the constrained Estimation of Distribution Algorithm EDA++.

Analysis of the components of the algorithm

• Algorithms

  • cmaes.m : COVARIANCE MATRIX ADAPTATION EVOLUTIONARY STRATEGY
  • geneticalgorithm.m : GENETIC ALGORITHM
  • mgd.m : CONSTRAINED ESTIMATION OF DISTRIBUTION ALGORITHM (Mixture of Gaussian Distribution)
  • pso.m : PARTICLE SWARM OPTIMIZATION

• InitialPop

  Contains initial seeds for comparing the algorithms. Seeds are generated for the benchmark of 13 problems.

• Problems

  The benchmark of 13 constrained continuous optimization problems.

• Results

  The obtained results for algorithms:

  • alg_mgd_lin_det : MGD + Linear Deterministic Method
  • alg_mgd_lin_sto : MGD + Linear Stochastic Method
  • alg_mgd_bis_det : MGD + Bisection Deterministic Method
  • alg_mgd_bis_sto : MGD + Bisection Stochastic Method
  • alg_cmaes : CMAES + Resampling Technique
  • alg_ga : GA + Penalty Function
  • alg_pso : PSO + Penalty Function

Comparison with other algorithms

• Algorithms

  • ALG_BPMAGES : BP-ϵMAg-ES
  • ALG_COLSHADE : COLSHADE
  • ALG_CORCO : CORCO
  • ALG_EDAPP : EDA++
  • ALG_ENMODE : EnMODE
  • ALG_LSHADE44 : LSHADE44

• CEC2020

  CEC 2020 real-world constrained optimization benchmark of 57 constrained continuous optimization problems.

• Solutions

  The obtained solutions for each algorithm regarding all problems in the benchmark (6 algorithms, 57 problems, 25 runs).

Credit

Abolfazl Shirazi

Zerua Tech S.L. | Basque Center for Applied Mathematics - BCAM | Universidad del Pais Vasco UPV/EHU

Email: abolfazl.shirazi@zerua.tech