Genetic Alg Flashcards

1
Q

Genetic Alg flow:

A

Initial population
Calc fitness
while termination cond not met{
select fitter individuals
recombine
mutate
evaluate fitness of all individuals
generate new population
}
return best individual

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
2
Q

Different types of selection:

A

Tournament selection (best of n)

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
3
Q

Why is tournament selection size population dependent?

A

If the tournament size = population, only best will be selected = converge = less exploring new areas

If tournament size = 1 = random selection

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
4
Q

What is a good crossover prob?

A

0.6-0.9 (promotes exploration of the search space)

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
5
Q

Different types of crossovers?

A

1 point 2 point uniform crossover

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
5
Q

between what and what should a crossover point be selected?

A

0-5

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
6
Q

Good mutation prob range?

A

0.01-0.1

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
7
Q

Why do we need crossover + mutation?

A

if there are only 00000 in both parents there is no chance to get a 1 from just cross over hence mutation

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
8
Q

Why do we need mutation?

A

To get out of a local optima.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
9
Q

What are the GA parameters?

A

Population size
selection type
crossover type/rate
mutation type/rate
stopping criteria (Num generations/runs or convergence

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
10
Q

Which part of the GA are low level heuristics?

A

Mutation + selection

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
11
Q

High level heuristics in GA?

A

Crossover

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
12
Q

Which heuristic performs exploration?

A

Mutation + random selection (even though random to a lesser extent)

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
13
Q

Which heuristic performs exploitation?

A

Crossover and fitness-Proportionate selection

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
14
Q

Different methods of selection:

A

Tournament
Fitness proportionate
Roulette-wheel
Elitist selection

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
15
Q

How does fitness proportionate selection work?

A

Calc total fitness
assign selection probability (divide each fitness score of each individual by the total fitness, this represents the proportion on the roulette wheel)

16
Q

How does uniform crossover work?

A

random subset is taken from parent 1, other bits from parent 2