Computational Intelligence

Lecture: Computational Intelligence - ZHAW
Lecturer: Dr. Carsten Franke
Author of Algorithms: Michael Kressibucher

Algorithms:

• Evolutionary Algorithms
  • Einkriterielle Evoulutionäre Algorithmen
    • Genetic Algorithm
    • Evolution Strategy
  • Mehrkriterielle Evolutionäre Algorithmen
    • Vector Evaluated Genetic Algorithm (VEGA)
• Metaheuristic
  • Simulated Annealing

API

Install Requirements

Terminalplot is required for all but simulated_annealing.py. It plots simple graphs within command line.

pip install -r requirements.txt

or

pip install terminalplot

Execute Scripts

Parameters of scripts can be changed in main function inside a script. Scripts then are executed in command line with python.

python evolution_strategy.py

python genetic_algorithm.py

python vector_evaluated_genetic_algorithm.py

python simulated_annealing.py

Evolutionary Algorithms

Problem solved with algorithms

Height of Cylinder h: 0 ≤ h ≤ 31
Diameter of Cylinder d: 0 ≤ d ≤ 31

Volume V:
Surface S:

By a genetic algorithm, diameter and height have to be determined, such that the surface of the cylinder is minimal.

Context Parameters

• Object Properties (Encoding of Problem):
• Strategy Parameters (Mutation, Recombination, Selection):
• Fitness:
• Size of Population at Time t:
• Number of Descendants: λ
• Number of Parents at recombination: ρ
• Max Number of Generations an Individual lives: κ
• k: index of individual

Genetic Algorithm

• Individuals
• μ = λ
• κ = 1
• constant

Binary Encoding

Phenotype: Individual, Creatur, Properties , l, u: lower and upper bound of interval of n parameters
Genotype: , (l: length of binary encoding)