Subject: | |
From: | |
Reply To: | |
Date: | Wed, 26 Oct 2011 01:58:44 +0200 |
Content-Type: | text/plain |
Parts/Attachments: |
|
|
Hello everyone!
Within my Computational Biology studies at the University of applied
siences Hagenberg (Austria), I am currently working on my bachelor
thesis. The topic is "Comparison of frameworks for genetic
algorithms".
I want to implement a simple sequential GA for a traveling salesman
problem. Based on my researches on ECJ, I came to the conclusion that
I will have to implement/create following parts:
- special integervector-individual, which has to contain unique genes
from 1 to its genomes length
- an special "Initializer"
- a "Problem" containing the evaluate() method, in which i evaluate
the length of the individuals path
- my own MutationBreedingPipeline and CrossoverBreedingPipeline, so no
illegal individuals are produced
- the parameter file
My questions are:
- Where should I put the distance Matrix? Should it be integrated in
the Problem class?
- Will i need anything else in addition to the implementation-ideas
above? (I know it is a stupid question, but ECJ is that huge, so i
feel totally lost about what already exists and what does not)
Thx for your time!
Best regards,
Matthias
|
|
|