hoh / gaproject

Project for the course of Evolutionary Computing

gaproject

Project for the course of Evolutionary Computing

How to run:

Get in the project's folder:

cd gaproject

Then run the gaproject directory:

python gaproject

How to run - distributed:

Launch the MongoDB database

Launch a few workers:

python gaprunner/worker

Launch an analyzer:

python analyzers/optimizer.py

gaproject's Issues

Modify basic example of TSP - Try OX crossover operator

Implement SCX crossover?

According to this article it should be better than ERX...

https://docs.google.com/viewer?url=http%3A%2F%2Fwww.cscjournals.org%2Fcsc%2Fmanuscript%2FJournals%2FIJBB%2Fvolume3%2FIssue6%2FIJBB-41.pdf

Implement loop removal

Choose termination condition

Implement new mutations

See:
Exchange Mutation: The exchange mutation operator selects two cities of the tour randomly and simply exchanges them. In various publications the exchange mu- tation operator is also referred to as swap mutation, point mutation, reciprocal exchange, or order-based mutation [LKM+99].
• Insertion Mutation: The insertion mutation operator [Mic92] randomly chooses a city, re- moves it from the tour and inserts it at a randomly selected place. An alternative naming for insertion mutation is position-based mutation
[LKM+ 99].
• Simple Inversion Mutation: The simple inversion mutation operator [Hol75], which is used in the TSP experiments of the book, randomly selects two cut points and sim- ply reverses the string between them.
• Inversion Mutation: The inversion mutation operator [Fog93] randomly selects a subtour, removes it, and inserts it in reverse order at a randomly chosen position. An alternative naming for inversion mutation is cut-inversion mutation [LKM+ 99].

Implement production of tables (csv?)

It should be possible to easily create tables in a final document.

Example of table

Implement loading of files

Créer des tests unitaires

Pour éviter les régressions.

Voir une doc sympa :

http://sametmax.com/se-simplifier-les-tests-python-avec-pytest/

Implement ERX crossover

Create custom mutation operator

Graph the stats

Use matplotlib to graph the stats of the process (diversity, best & worse individual over time, ...)

To be integrated GNX crossover

http://biopython.org/DIST/docs/api/Bio.GA.Crossover.GeneralPoint-pysrc.html

Improvement in sample mechanisms?

Simple tournament to select the parents that mate in each interaction. See exercise 3 of genetic algos part 1