Algorithms - Binary and Real Genetic Algorithms

The OSTRICH genetic optimization algorithms

Initial Publication:

Modified:

The following optional group will configure either the binary- or real-coded genetic algorithm and will be processed if [ProgramType] is set to “BinaryGeneticAlgorithm” (i.e. BGA) or “GeneticAlgorithm” (i.e. RGA).

BeginGeneticAlg
ParallelMethod       [pmethod]
PopulationSize       [pop_size]
MutationRate         [mut_rate]
Survivors            [nelites]
NumGenerations       [numgens]
InitPopulationMethod [imethod]
ConvergenceVal       [conv_val]
EndGeneticAlg
BeginGeneticAlg
ParallelMethod       synchronous
PopulationSize       50
MutationRate         0.05
Survivors            1
NumGenerations       10
InitPopulationMethod random
ConvergenceVal       1.00E-4
EndGeneticAlg

Figure 1: General Format (left) and Example (right) of the Binary and Real-coded Genetic Algorithm Groups

Where BeginGeneticAlg and EndGeneticAlg are parsing tags that wrap a set of algorithm configuration variables. These variables are described below:

  • ParallelMethod: This variable controls how the algorithm is to be run if parallel computing is used. Supported values are “synchronous” and “asynchronous”. The default value is “synchronous”.
  • InitPopulationMethod: This variable controls how the algorithm configures the initial population of candidate solutions. Supported values are: “random”, “LHS” (Latin Hypercube Sampling), and “QuadTree”. The default value is “random”.
  • PopulationSize: The population size. The default value is 50.
  • MutationRate: The mutation rate for child members. The default value is 0.05 (i.e. 5%).
  • Survivors: The number of elites who pass unchanged to next generation. The default value is 1.
  • NumGenerations: The number of generations in the search algorithm. The default value is 10.
  • ConvergenceVal: This is the convergence value for the algorithm. If the relative difference between the current minimum and the median of the latest generation is less than or equal to this value, the algorithm will halt. The default value is 1.00E-4.

References

Chan Hilton, A. B.,Culver, T. B. 2000. Constraint handling for genetic algorithms in optimal remediation design. Journal of Water Resources Planning and Management 126, 128-137.

Yoon, J.-H.,Shoemaker, C. A. 2001. Improved real-coded GA for groundwater bioremediation. Journal of Computing in Civil Engineering 15, 224-231.

Yoon, J.-H.,Shoemaker, C. A. 1999. Comparison of optimization methods for ground-water bioremediation. Journal of Water Resources Planning and Management