Forums before death by AOL, social media and spammers... "We can't have nice things"
|    comp.ai.fuzzy    |    Fuzzy logic... all warm and fuzzy-like    |    1,275 messages    |
[   << oldest   |   < older   |   list   |   newer >   |   newest >>   ]
|    Message 491 of 1,275    |
|    Walter Banks to Ala Qumsieh    |
|    Re: Is Anyone Actually Doing Anything In    |
|    22 Jun 05 08:05:15    |
      From: walter@bytecraft.com              The place to start is to write down all of the individual choices and       why just as you have here. (My preferred method is to use a       spreadsheet as an electronic blackboard)              From that list make a list of all of the choices you have to make.       (consequences)              Make a list of all of the things that affect the consequences and why.       This will form the basis of the Linguistic variables.              Start writing down how linguistic variables affect consequences, this       will be the basis of fuzzy logic rules.              Group the fuzzy logic rules by consequence. This will start the       organization of fuzzy functions.              Fuzzy functions are the evaluation of competing rules to identify       the best overall outcome.              Just a place to start.              w.                                   Ala Qumsieh wrote:              > EarlCox wrote:       >       > > BUT no one is discussing problems with the actual use       > > of fuzzy logic in real world applications.       >       > I have been thinking quite a bit about a certain application related to       > my job. Note however that I haven't done any implementation yet. The       > main reason for this is that I'm not sure how to start. I'm sure there       > are other in my shoes. Perhaps someone can put me (and others) on the       > right path here.       >       > I am an ASIC designer. One of the problems we face is where to       > physically place large macros and RAM modules within a block so as to       > minimize routing congestion and timing violations. There are various       > constraints, mainly:       >       > - The connections to the RAMs and macros. Those are specified in the       > form of a netlist. The netlist can be simplified to prune out any       > non-critical logic, and to simplify the connectivity.       >       > - The block's geometry. This is typically rectangular, but could also be       > piecewise rectangular.       >       > - Clock constraints. Some clocks are more critical (ie faster) than others.       >       > I'm only interested in the macro and RAM placement, not any other logic.       > The results don't need to be perfect.       >       > Is FL the correct approach? Do I need to use NNs to optimize my       > termsets? Or are GAs a better choice? How do I choose my variables rule       > set? I know that a lot of it is fuzzy (pun intended), but I'm not sure       > how to start.       >       > --Ala              --- SoupGate-Win32 v1.05        * Origin: you cannot sedate... all the things you hate (1:229/2)    |
[   << oldest   |   < older   |   list   |   newer >   |   newest >>   ]
(c) 1994, bbs@darkrealms.ca