| |
Hi all, a simple question ( I hope)
I'm running a small GA on my rcx block , and I'm fairly happy qwith my
results so far, it evolves well tot he conditions, however I'd like to try
something new <g>...
I'm going to try to evolve sub-sumptive architecture. To do this I'd like
to evolve an 'instinctual' layer and a more 'cogniscant' layer. To do this
I'd like to
re-write my GA in a more object-orientated fashion. Now I know C++ can be
used with the legOS system, but I'd like to know how much of a memory/speed
over head this would impose on my code, as both these things are important
to me
Any help would be gladly received :)
- Ciaran
|
|
Message has 3 Replies:
 | | Re: C++ OverHead
|
| I seem to recall someone doing some testing that showed that C++ programs worked out to take out somewhere around 2 to 3 times the space of comparable C programs. However, I'm sure that varies. Luis (...) ---...--- "Nobody ever said that democracy (...) (24 years ago, 23-Jan-01, to lugnet.robotics.rcx.legos)
| | | Re: C++ OverHead
|
| (...) Hi, unfortunately I am unable to answer your question, however I am very interested in what you are doing with GA. I recently recieved my RCX and have done GA work in the past as part of my schoolwork. I would love to try using some GAs with (...) (24 years ago, 24-Jan-01, to lugnet.robotics.rcx.legos)
| | | Re: C++ OverHead
|
| Hi, The answer here depends on your coding style. If you are using non-derived concrete classes, with simple member data and the default constructor/destructor, then it costs ~10-20 bytes/class. If you are using non-derived concrete classes, with (...) (24 years ago, 24-Jan-01, to lugnet.robotics.rcx.legos)
|
4 Messages in This Thread:
 
- Entire Thread on One Page:
- Nested:
All | Brief | Compact | Dots
Linear:
All | Brief | Compact
|
|
|
|
