Can I get better results by using the final population from a previous run as the initial population for a new run in the genetic algorithm?
I think, answer may also be in your this query, if you kindly check it according to your query.
If you take the final population as is, then it will be the same as continuing the previous run. So it will not make any sense.
If you mean selecting some individuals from the final population and fill the rest with random solutions then it may converge faster than using a totally random initial population. However, this time, you take the risk of getting stuck at local optima because of the lack of diversity. Actually, the second strategy I mentioned here called elitism that is a common strategy in GA and guarantees to keep a predefined number of good solutions at each iteration.
this is based on the selection a new function to fit the chromosome or i think you have to learn you data and this may be needed to more time
If run means one generation - absolutely yes. It is very desirable also to use elite selection among the others.
- Badges
- Science topic
- Similar topics
- Biological Science
- Kinesiology
- Running