Genetic Algorithms

Basic Genetic Algorithm – Flow Chart

1. Initial Population

1. Initial Population

ON

ON

| | GENERATE RANDOM POPULATION (POSSIBLE SOLUTIONS) | 2. Fitness Evaluation

2. Fitness Evaluation

3. Selection

3. Selection

| | EVALUATE THE FITNESS OF EACH (BASED ON THE FITNESS FUNCTION) | | | CHOOSE PARENT FACTORS (BETTER FITNESS = BETTER CHANCE) | 4. Crossover

4. Crossover

5. Mutation

5. Mutation

| | CROSSOVER THE PARENT TRAITS TO FORM NEW CHILDREN. (PROBABILITY) |

| | MUTATION PROBABILITY APPLIED (MAINTAINS GENETIC DIVERSITY) | Acceptable?
Acceptable?
| | IF OPTIMIZATION CONDITIONS ARE NOT MET(REPEAT STEPS 2-5) * OR | Yes

Yes

End Process
End Process | | IF THE MAXIMUM GENERATIONS ARE MET (TERMINATE) * OR | | | IF SATISFACTORY FITNESS LEVEL IS REACHED (END THE PROCESS) |

KEY TERMS * INDIVIDUAL
Any possible solution to the problem at hand, usually expressed in binary code * POPULATION
Group of all individuals * CHROMOSOME
Blueprint for an individual usually expressed in binary code. (Ex: 011011) * GENE
An individual value in a chromosome, usually expressed as a “1” or “0” * PARENTS
An original “individual” solution in the GA process that has passed the fitness function * CHILDREN
A new solution to the problem formed through crossover and mutation from the parent solutions * SEARCH SPACE
All possible solutions to the problem at hand and every possible place for the algorithm to look * SELECTION MECHANISM
How you choose the parent solutions that will in turn create child solutions, and iteratively optimize the problem * FITNESS FUNCTION
A function that describes how good or bad a solution is * CROSSOVER
Recombine parents at a break point in the chromosome to form children * MUTATION
Invert individual genes at a specified probability to promote genetic diversity