Evolutionary Algorithms Flashcards
What does an evolutionary algorithm provide?
It provides an iterative solution which evolves a population of individuals
What does each individual of a population contain?
It contains a potential solution to the problem.
How does each individual get tested to see how good its solution is?
it is evaluated by a fitness function, which measures the quality of its potential solution.
What is a chromosome in a genetic algorithm?
one of the members of the population.
What is a parent in a GA?
A chromosome that has had children generated from it.
What does a generation refer to in GA’s?
The number of iterations through the algorithm.
Why do we need to find a balance between exploratory population members and exploitative members?
- Exploratory to look around the search space.
- Keep track of the potential solutions we think are good.
What does rank based selection do?
- Sorts individuals according to their fitness
- Selects individuals with the probability proportional to rank only
What are two techniques we use to generate children?
Crossover and Mutation
What happens in crossover?
We choose a point in both parents, then children are generated from both sides of that point. e.g:
P1(1010) C1(1001)
P2(0101) C2(0110)
What does mutation do?
Makes local changes to a chromosome, based on probability of a bit (or whatever data in chromosome) has of changing.
Probability of this is defined by the developer.
How does mutation affect the chromosome?
It makes the chromosome more exploratory.
How are crossover and mutation applied to a population?
Applied based on chance.
What is the chance of crossover happening?
Fairly high, to ensure that we focus on a potential solution.
