Setting parameters in the Genetic algorithm
The default parameters in GA should be adequate for most problems. However, if the user needs to explore different methods, the following parameters may be set for GA.
| Option |
Parameter
|
| Population Size | Population size |
| Number of Generations | Number of generations |
| Selection Operator | Selection operator: Tourament, Roullette, SUS |
| Tournament Size |
Tournament size for tournament selection operator
|
| Elitism |
Switch elitism for single objective GA
|
| Number of Elites |
Number of elites passed to next generation
|
| Encoding variable |
Type of encoding for a variable: Binary=1, Real=2
|
| Numbits variable |
Number of bits assigned to a binary variable
|
| Crossover type |
Type of real crossover: SBX, BLX
|
|
Crossover probability
|
Real crossover probability
|
| Alpha value for BLX | Value of α for BLX operator |
| Crossover distribution |
Distribution index for SBX crossover operator
|
| Mutation probability |
Mutation probability in real-space
|
| Mutation distribution |
Distribution index for mutation operator
|
| Algorithm Subtype | Mulit-objective optimization algorithm: NSGA II, SPEA II |
|
Restart Interval
|
Frequency of writing restart file for direct GA For multi-objective problems, this parameter governs the frequency of writing TradeOff files
|
| Max Repeat Optimum/Generations | Maximum number of generations allowed to repeat as a fraction of the totla number of generations allowed. |
|
Constraint Handling
|
Constraint handling types: Deb Efficient Constraint Handling, Penalty
|