Unit 3: Genetic Algorithms – Probable
Semester Questions and Answers
1. Concept of "Genetics" and "Evolution" and its Application to
Probabilistic Search Techniques
Genetics and evolution in nature refer to the biological processes of inheritance and natural
selection.
In computational intelligence, Genetic Algorithms (GAs) mimic these natural processes to
perform search and optimization tasks.
- Genetics provides the metaphor of encoding potential solutions as "chromosomes."
- Evolution refers to selection, crossover, and mutation to improve generations of solutions
over time.
- GAs belong to a class of evolutionary algorithms that perform probabilistic search, not
deterministic.
Applications: Traveling Salesman Problem, Neural Network training, Job Scheduling, Game
Playing, Function Optimization.
MCQs:
1. Genetic Algorithms are based on:
A) Deterministic search
B) Dynamic programming
C) Biological evolution principles ✅
D) Random sampling only
2. In a GA, a chromosome represents:
A) An organism
B) A potential solution ✅
C) A cell
D) A random guess
2. Basic GA Framework and Different GA Architectures
GA Framework:
1. Initialization: Generate a random population of individuals (chromosomes).
2. Evaluation: Calculate fitness of each individual.
3. Selection: Select individuals based on fitness.
4. Crossover: Combine parts of two parents to produce offspring.
5. Mutation: Randomly alter genes to maintain diversity.
6. Replacement: Replace old population with new offspring.
7. Termination: Stop when convergence or generation limit is reached.
Different GA Architectures:
- Generational GA: Replaces entire population each generation.
- Steady-State GA: Replaces few individuals at a time.
- Parallel GA: Population divided into subgroups to process in parallel.
- Distributed GA: Use of isolated subpopulations (islands) with occasional migration.
MCQs:
1. What is the first step in a basic Genetic Algorithm framework?
A) Mutation
B) Initialization ✅
C) Selection
D) Fitness Evaluation
2. Steady-State GA differs from Generational GA because:
A) It uses more chromosomes.
B) It never stops.
C) Only a few individuals are replaced at each step. ✅
D) It does not use crossover.
3. GA Operators: Encoding, Crossover, Selection, Mutation, etc.
Encoding: How solutions are represented.
- Binary Encoding: e.g., 101101
- Real-valued Encoding: e.g., 3.4, 5.2
- Permutation Encoding: For order-based problems (e.g., TSP)
Selection: Chooses fitter individuals for reproduction.
- Roulette Wheel Selection
- Tournament Selection
- Rank Selection
Crossover (Recombination): Combines two parents to produce offspring.
- Single Point
- Two Point
- Uniform Crossover
Mutation: Randomly alters genes.
- Bit Flip (Binary)
- Swap Mutation (Permutation)
- Gaussian Mutation (Real-valued)
Elitism: Preserving best individuals from one generation to the next.
MCQs:
1. In GA, which operator helps maintain genetic diversity?
A) Selection
B) Crossover
C) Mutation ✅
D) Encoding
2. Which selection technique is based on random chance but weighted by fitness?
A) Tournament Selection
B) Rank Selection
C) Roulette Wheel Selection ✅
D) Linear Scaling
4. Solving Single-Objective Optimization Problems using GAs
A single-objective optimization problem aims to either maximize or minimize a given
function.
Steps using GA:
1. Define the objective function f(x)
2. Encode the solution into chromosomes
3. Initialize population randomly
4. Evaluate fitness for each chromosome using f(x)
5. Apply GA operators: selection, crossover, mutation
6. Repeat until stopping condition is met
Example:
Maximize f(x) = x^2 where 0 ≤ x ≤ 31
- Binary encoding: 5 bits
- Use selection, crossover, and mutation to evolve better values of x over generations.
MCQs:
1. In single-objective optimization, the goal is to:
A) Satisfy all constraints
B) Maximize or minimize one function ✅
C) Minimize multiple costs
D) Generate all possible solutions
2. What is the fitness of a solution in GA?
A) Its complexity
B) Its execution time
C) A value representing solution quality ✅
D) Random number
