Binary Genetic Algorithms

1
Binary Genetic Algorithms

• Chapter 2

2
Recall - How does a GA work?

• 1. Define cost function, variables, constraints
• 2. Define decoding
• 3. Create Initial Population
• 4. Select parents
• 5. Do crossover
• 6. Perform Mutation
• 7. Select Survivors

3
Binary Encoding

• Use binary to represent each number
• Problem quantization error

quantized function
function
4
Binary Encoding

• Do random sampling (initial population)
• Find the interval for each sample
• Take the average of the high and low of the
  interval of each sample and assign the average as
  the sample's value

5
Selection

• Sort parents by weight and select them in order
• Select parents randomly
• Rank weighting
• rank parents
• pi (N 1 - i)/ (sum of ranks) (e.g. linear
  ranking with s0)
• exponential ranking
• Cost weighting
• e.g. ci ci - clowest
• pi ci / (sum of costs) - roulette wheel
• stochastic universal sampling
• Tournament Selection
• select few elements at random
• pick best out of them

6
Recombination (mating)

• one point crossover
• two point crossover
• uniform

7
Mutations

• randomly mutate one or more bits
• no mutation on final solution
• recall Gray encoding

8
Survivor Selection

• age-based
• fitness-based
• recall elitism / genetor