genetic algorithm discrete variable

[ziedhosni]

Hi everybody,

I am using a genetic algorithm VI to optimize a chemical process and I am using contineous variables to define my inputs. Now, I would like to change the type of inputs to discrete variables ( for example, the rotation per minute in the reactor can take juste few value: 200 500 700 1000).

I found in some papers we can use an hybridization of contineous and dicrete variables to define the inputs. If someone can help me to sort out this problem of integrating discrete variables in the GA and if someone has the experience, what are the drawbacks of using discrete variables with the GA.(if you can help me with references or example, what I found is old and not well explained).

Thank you a lot.

[ShishKeBobby]

Hi Zierdhosni,

So from what I can tell, you're interested in making a hybrid continuous/discrete variable?   Could you post a link to the reference material you mentioned, if there is one?  Thanks.

[ziedhosni]

Hi, You find attached some papers I found. But they are hard to understand as I am a chemical engineer.

[ShishKeBobby]

I agree, this is a very complex issue.  I would reccomend you break down your problem into small parts then repost them individually.  As of right now it's a lot to chew on!