Human evolutionary model: A new approach to optimization

  • Authors:

  • Oscar Montiel;Oscar Castillo;Patricia Melin;Antonio Rodríguez Díaz;Roberto Sepúlveda

  • Affiliations:

  • CITEDI-IPN, Av. del Parque #1310, Mesa de Otay Tijuana, BC, Mexico;Department of Computer Science, Tijuana Institute of Technology, P.O. Box 4207, Chula Vista, CA 91909, USA;Department of Computer Science, Tijuana Institute of Technology, P.O. Box 4207, Chula Vista, CA 91909, USA;FCQI-UABC Tijuana, BC, Mexico;CITEDI-IPN, Av. del Parque #1310, Mesa de Otay Tijuana, BC, Mexico

  • Venue:
  • Information Sciences: an International Journal
  • Year:
  • 2007

Quantified Score

Hi-index 0.07

Visualization

Abstract

The aim of this paper is to propose the Human Evolutionary Model (HEM) as a novel computational method for solving search and optimization problems with single or multiple objectives. HEM is an intelligent evolutionary optimization method that uses consensus knowledge from experts with the aim of inferring the most suitable parameters to achieve the evolution in an intelligent way. HEM is able to handle experts' knowledge disagreements by the use of a novel concept called Mediative Fuzzy Logic (MFL). The effectiveness of this computational method is demonstrated through several experiments that were performed using classical test functions as well as composite test functions. We are comparing our results against the results obtained with the Genetic Algorithm of the Matlab's Toolbox, Evolution Strategy with Covariance Matrix Adaptation (CMA-ES), Particle Swarm Optimizer (PSO), Cooperative PSO (CPSO), G3 model with PCX crossover (G3-PCX), Differential Evolution (DE), and Comprehensive Learning PSO (CLPSO). The results obtained using HEM outperforms the results obtained using the abovementioned optimization methods.