Difference between revisions of "Cellular genetic algorithm (analysis)"

Latest revision as of 11:15, 31 May 2013

Analysis title

Cellular genetic algorithm

Provider

Institute of Systems Biology

Class

MOCellOptMethod

Plugin

ru.biosoft.analysis.optimization (Common methods of data optimization analysis plug-in)

[edit] Multi-objective cellular genetic algorithm (MOCell)

The cellular model of the genetic algorithms works on a single population of candidate solutions (individuals)¹. MOCell is an adaptation of the algorithm to the multi-objective field². The main steps of the algorithm are the following.

Initialization. The one individual in the initial population is defined by the user, the others are uniformly randomly generated.

Evaluation. Once we initialized the population, or when a new solution offspring is created, it is necessary to calculate the fitness values for all individuals in this population. For this purpose, we need to find values of objective and penalty functions. (The latter is in the case of the constraint optimization.)

Selection. Through the selection, we favor solutions, which have the highest fitness value (that means the lowest values of both objective and penalty function values) in the search.

Recombination. The recombination combines two parent solutions for creating a new offspring using Simulated Binary Crossover (SBX).

Mutation. The mutation randomly modifies the offspring using the polynomial operator.

Replacement. Individuals of the offspring population created through selection, recombination and mutation, replace individuals of the parent population if their fitness values are higher.

[edit] References

1. E Alba and B Dorronsoro. "Cellular Genetic Algorithms". New York: Springer, 2008.
2. AJ Nebro, JJ Durillo, F Luna, B Dorronsoro and E Alba. "A Cellular Genetic Algorithm for Multiobjective Optimization". International Journal of Intelligent Systems, 24(7): 726-746, 2009.