Parameter Settings For Genetic Algorithms

A typical genetic algorithm contains the following parameters:

Initilization Parameters

Population Size How many individuals in a population
Initilization Methods
  • Random initilizing: Individuals are randomly generated within the boundaries.
  • Uniformly initilizaing: Individuals are even spreaded within the boundaries.
Random Seed Random seed  initilization method. A random number will depend on the random seed. there are three way to set the random seed:
  • Randomly. Each run, the random seed is changing.
  • Fixed. the random seed is a fixed number. you can repeat the same process and obatin the same results.
  • Change with the run order. The random seed is set to the current run order.
Random Methods Setting of random number generating method.
Distance Method There are three methods used in this system:
  • Euclidean distance
  • Different between variables
  • User defined method. This shold be included in your model
Remove infeasible individial If an individual is not in the feasible region, it will be removed from a population and an existed invidial will be copied into the population
Including the initial point The initial point will be added into the popultaion.

Selection Parameters

Method The method of selecting individuals to generate a new generation
Scaling Method for scaling the fitness of each individuals
Order Method After Selection process, this method is used to order the individuals in the population

Crossover Parameters

Method Crossover method
Probability Crossover probability
Select Parent Method of selecting paraents for crossover
Children Method for placing childrens
Maximum Distance The maximum distance of paraents. If it is 0, the maximum distance is infinite.

Mutation Parameters

Method Mutation methods
Probability Mutation probability
Select Parent Method of selecting paraents for crossover
Children Method for placing childrens
Neighbohood It is value between 0 and 1.

Termination (Stop) Parameters

Maximum Genetation Tthe maximum number of generations.
Minimum generation The number of generation must be greater than this number.
Species Number The minimum species number
Solutions The minimum number of solution.
Evluation of objectives The maximum evaluation of objectives.

Related Pages: