Heterogeneous Strategy and its Application in an Improved Micro-Genetic Algorithm

Abstract:

Heterogeneous strategy, population isolation, arithmetic crossover and optimum reserved strategy are used to improve micro-genetic algorithm (mGA) in this paper. Heterogeneous strategy is used to improve the probability of convergence to global optimal solution and quicken up the convergence. Reset frequency is decreased while the global and local searching capabilities of mGA between two resets are enhanced, which makes mGA searching the parameter space intelligently as the mode recognition information is preserved as much as possible. Adaptive random mutation, which used existing genetic information of the current groups, is used to increase efficient search. Finally, standard functions testing demonstrate that the improved mGA can find better optimum solutions with less computing cost than standard genetic algorithm (SGA).

• Periodical: Advanced Materials Research (Volumes 542-543)
• Pages: 1467-1470
• DOI: https://doi.org/10.4028/www.scientific.net/AMR.542-543.1467
• Online since: June 2012
• Authors: Le Wei Yan, Yang Yang Chen
• Keywords: Adaptive Random Mutation, Arithmetic Cross, Heterogeneous Strategy, Micro Genetic Algorithm, Real Code

References

[1] Marinakis Y, Marinaki, M. A hybrid genetic-particle swarm optimization algorithm for the vehicle routing problem . Expert Systems with Applications, 37(2): 1446-1455 (2010).

DOI: 10.1016/j.eswa.2009.06.085

[2] Amin J, Mohammad A S, Reza T M. A hybrid algorithm based on particle swarm optimization and simulated annealing for a periodic job shop scheduling problem [J]. International Journal of Advanced Manufacturing Technology, 9:1-14 (2010).

DOI: 10.1007/s00170-010-2932-8

[3] Krishnakumar K. Micro-genetic algorithm for stationary and non-stationary function optimization [M]. SPIE Proceedings: Intelligent Control and Adaptive Systems, 289-296 (1989).

DOI: 10.1117/12.969927

[4] Hamzeh A, Rahmani A. Approximating arbitrary reinforcement signal by learning classifier systems using micro genetic algorithm [J]. Fundamenta Informaticae, 86(1-2): 93-111 (2008).

[5] Gerald W., Remegio C.J., Stephen M. G., Rolf F.,Shawna G., Jeffrey J. S., George W. M., and Gary M. B. A hybrid genetic algorithm for multi－objective problems with activity analysis-based local search. European Journal of Operational Research.193,195-203 (2009).

[6] Pei C.C., Shih H.C., and Mani V. A hybrid genetic algorithm with dominance properties for single machine scheduling with dependent penalties. Applied Mathematic Modelling. 33, 579-596 (2009).

DOI: 10.1016/j.apm.2008.01.006