Evolutionary AI wk 3

What is representation in evolutionary algorithms?

The way an individual (solution) is encoded so it can be manipulated by evolutionary operators.

Why is choosing a representation important in evolutionary algorithms?

Because it determines how mutation, crossover, and evaluation can be applied, and whether generated individuals remain valid solutions.

What is the difference between genotype and phenotype?

The genotype is the encoded representation; the phenotype is the actual solution in the problem space.

What is binary representation?

Encoding individuals as bitstrings composed of 0s and 1s.

When is binary representation typically used?

For problems where parameters can be naturally discretized or represented by fixed-length binary codes.

What are common mutation operators for binary representation?

Bit-flip mutation—independently switching bits from 0→1 or 1→0.

What is a drawback of binary representation?

It may require extra encoding/decoding steps for real-valued problems.

What is integer representation?

Individuals are represented as vectors of integers.

When is integer representation appropriate?

When the problem variables take discrete values (e.g., categories, counts, indices).

What kind of mutation is used for integer representations?

Randomly replacing a gene with an integer sampled within the allowed bounds.

What is floating-point representation?

Individuals are represented as vectors of real numbers.

When should floating-point representation be used?

In continuous optimization problems where decision variables are real-valued.

What is a common mutation operator for real-valued representation?

Adding Gaussian (normal) noise to each variable.

What is an advantage of floating-point representation?

It avoids encoding overhead and often leads to smoother optimization.

What is permutation representation?

Individuals are represented as an ordered list (a permutation) of elements.

For which problems is permutation representation necessary?

Ordering problems like the Travelling Salesman Problem (TSP) or scheduling.

Why can't standard crossover be used with permutation representation?

It may create invalid permutations with repeated or missing elements.

What type of mutation is used in permutation representations?

Swap or insertion mutation—changing the order of elements.

What determines the best representation to use?

The structure of the problem (continuous, discrete, or ordering).

What is the main goal when selecting a representation?

Ensuring solutions remain valid and that variation operators work effectively.