Please enable JavaScript.
Coggle requires JavaScript to display documents.
Evolutionary Computing (Components (Population (Holds the candidate
…
Evolutionary Computing
-
2 Competing
Forces
- Increasing population
- Diversity by genetic operator
- Crossover, Mutation
- Push toward novelty (the quality of being new)
- Decreasing population
- Diversity by selection
- Parent, Survivor
- Push toward quality
Components
Representation
Level
Genotype
- Code to denote the object
- The inside
- Chromosome
Phenotype
- Object in original problem context,
- The outside
Mapping
Encode
- Phenotype -> Genotype
- Not necessary 1-to-1
Decode
- Genotype -> Phenotype
- Must be 1-to-1
-
Population
-
-
-
-
Diversity
- Number of different fitnesses/individual present
Selection
-
-
-
Probabilistic
- High quality solutions more like to be selected
- Not guarantee
- Worst solution has non-zero probability to be selected
Variation Operator
- Arity 1 : mutation
- Arity >1 : recombination
- Arity 2: crossover
- Airty >2 : seldom used in EC
Mutation
Cause small, random variance
-
-
-