Re: GAs and TSP-Type Problems

"Steve McGrew" <[EMAIL PROTECTED]> wrote:

> "Kent Paul Dolan" <[EMAIL PROTECTED]> wrote:

[whatever]

> I've mentioned Generator, and there are also
> several others that allow the user to define the
> problem in an Excel spreadsheet.  

I don't mean to pick a quarrel, but you sure seem
to want one.

As bad taste as flogging my tinker toy freeware
GAs as programming examples might be, flogging a
commercial product repeatedly in a technical
newsgroup rates horse whipping.

An "Excel" spreadsheet solution might be dandy for
tinker toy problems, and I count some pretty
damned impressive ones as "tinker toy", but even
mine tend to use on the order of a hundred
different approaches and tens of thousands of
lines of code to solve one single instance of one
easily described kind of problem that happens also
to be incredibly difficult, and Microsoft has never
made nor sold a product sufficient to withstand
that kind of assault in its existence.  "Generator"
is not the all-purpose answer to every GA query, so
stop offering it as such, while you still have
Net connectivity.

xanthian.

And OBTW, "Traveller" for the first time
successfully passed a trillion city to city
distance calculations without me nuking it by
accident while playing NetHack or something
equally bogus, a milestone meaningless except for
the ego boost it gave me (I'm double spacing the
usual Traveller status window display to cater for
expected browser mangling, mine and yours; by the
time Mozilla's bug adding extra space around
quoted lines had done it's mischief, these might
be as much as quadruple spaced, but quoting the
lines (usually) protects them from folding at my
end), though it does seem to prove that it is
possible for a really complex Java program to run
for a long time (a bit over two weeks here) under
WinXP SP1, sucking down a deliberately metered 95%
of the CPU cycles and modestly doing other chores
with the remainder of the resources without
crashing the dappy computer OS if one is very,
very careful:

==LOG STATUS== TravellerStatus Display Contents:

> current FB genome by: RollingCrossover

> current first best (FB) genome ID: 4; clones: 1; same fitnesses: 1; unique 
> fitnesses: 6 / 21

> brute force solution time needed for 1,248 nodes at 2000 MegaFLOPS: 5.8539*10^3307 
> years.

> candidates -- improved / annealed / tabued / mutated / considered

> overall: 8,174 / 0 / 0 / 1,389 / 267,720, last generation: 0 / 0 / 0 / 0 / 20

> below value meanings -- best / at last improvement / currently / at start / worst 
> ever; other stuff

> shortest tour: 10,485.8894 / 10,485 / 10,485.8894 / 332,906 / 332,906

> median tour: 10,486.1147 / 10,486 / 10,486.1147 / 343,419 / 343,419

> average tour: 10,486.1017 / 10,534 / 10,487.6436 / 343,303 / 343,303

> longest tour: 10,486.1147 / 11,288 / 10,500.9254 / 354,975 / 354,975

> unique edges count: 1,255 / 1,313 / 1,286 / 25,831 / 25,831; playfield size, x/y: 
> 600/600

> unique edges total length: 10,574.9616 / 12,272.992 / 10,964.8234 / 7,107,639.805 / 
> 7,107,639.805

> unique edges average length: 8.4177 / 9.3473 / 8.5263 / 275.1593 / 275.1593

> permute: (limits::whiffs) now/highest::inARow/limitBumpsAt 6 / 8 :: 5090 / 40380; 
> bumps on

> generations at last: improvement / generation ; # of improvements: 10,268 / 13,386 # 
> 118

> elapsed seconds for: setup / run at last: improvement / generation 4 / 722,075 / 
> 1,207,091

> lengths taken at: setup / last improvement / last generation: 52,039 / 
> 568,852,840,487 / 1,232,479,048,524

> annealing not enabled; tabuing not enabled



-- 
Posted via Mailgate.ORG Server - http://www.Mailgate.ORG