/Derandomized_GA

Primary LanguagePython

Derandomized_GA

TODO:

Data Generation:

  1. Create Data set
  2. https://people.sc.fsu.edu/~jburkardt/datasets/tsp/tsp.html https://github.com/ezstoltz/genetic-algorithm/blob/master/genetic_algorithm_TSP.ipynb

Standard GA: RAZ - ✅

  1. Initialization
  2. Selection - Tournament
  3. Mutation - flip 2 indices
  4. XO
  5. Fitness

Embedding: SNIR

  1. Turn Individual into a compress version with Random Walk.

Neural Network: SNIR

  1. Grind K mutations on an individual.
  2. Choose if to insert bad examples... \ if no improvment target vector will be the zero vector
  3. Train NN, use softmax for prediction which indices to flip

Visualization & Results: AMIR

  1. Run experiments with different hyper paramters.
  2. create nice plots with results
  3. search for p-value with permutation testing.