I worked up something 20 or more years ago, might have been pretty
domain-specific (optimising changes in diet!!!) and was in Dyalog APL.
I've probably still got it. and could look at ways of making a readable
text-file, or workspace if you've got Dyalog. It worked on the
Lagrangian, I think
Has anybody written a quadratic optimization solver in J? Or is there one in
any of the packages?
Examples: https://en.m.wikipedia.org/wiki/Quadratic_programming
--
For information about J forums see http://www.jsoftware.com/f
Looks very concise
We were out for Liz's birthday, so I hadn't thought about it!
Mike
On 12/03/2016 23:44, Rob Hodgkinson wrote:
Agree, very nice solution Raul.
On 13 Mar 2016, at 10:02 AM, Kip Murray wrote:
Very nice, Raul! Much shorter than my Rube Goldberg approach. --Kip
On Sa
Agree, very nice solution Raul.
> On 13 Mar 2016, at 10:02 AM, Kip Murray wrote:
>
> Very nice, Raul! Much shorter than my Rube Goldberg approach. --Kip
>
> On Saturday, March 12, 2016, Raul Miller wrote:
>
>> cv=: ([: (+`%/) 1 }.,)\@|:
>>
>> I hope this helps...
>>
>> --
>> Raul
>>
>>
Very nice, Raul! Much shorter than my Rube Goldberg approach. --Kip
On Saturday, March 12, 2016, Raul Miller wrote:
>cv=: ([: (+`%/) 1 }.,)\@|:
>
> I hope this helps...
>
> --
> Raul
>
>
> On Sat, Mar 12, 2016 at 1:21 PM, Kip Murray > wrote:
> > Here you go:
> >
> > nume =: 1 , 1 % 4x
cv=: ([: (+`%/) 1 }.,)\@|:
I hope this helps...
--
Raul
On Sat, Mar 12, 2016 at 1:21 PM, Kip Murray wrote:
> Here you go:
>
> nume =: 1 , 1 % 4x * _1 + 4 * [: *:@>:@i. <:
>
> dene =: 1 1r2 , 1 $~ _2 + ]
>
> I think I got those from Abramowitz and Stegun.
>
>(_1 , nume 6),: dene
Here you go:
nume =: 1 , 1 % 4x * _1 + 4 * [: *:@>:@i. <:
dene =: 1 1r2 , 1 $~ _2 + ]
I think I got those from Abramowitz and Stegun.
(_1 , nume 6),: dene 7
_1 1 1r12 1r60 1r140 1r252 1r396
1 1r211 1 1 1
--Kip
On Saturday, March 12, 2016, Raul Miller wro
How do you compute the first two rows?
Thanks,
--
Raul
On Saturday, March 12, 2016, Kip Murray wrote:
> The challenge is at the end. First a table for a finite continued fraction
> that approximates e =: ^ 1 .
> --Kip Murray
>
>
> The table below summarizes a finite continued fraction which
Thanks, Rob! This was exactly what I was looking for.
On Sat, Mar 12, 2016 at 12:37 AM, Rob Hodgkinson wrote:
> Hi Devon, try ‘stats’ add-on in Package Manager…
>
>load 'stats/base/distribution'
>load 'plot'
>plot binomialdist 0.8 20
>plot binomialdist 0.2 20
>
> HTH, Rob
>
> >
What you are doing is very nice.
I call mine a point direction approach. A point on a straight line is
described by
S + t * D
Where t is a real number, and S And D are points in a vector space, the
starting point and the direction point.
Another approach valid in any vector space is the two-po
The challenge is at the end. First a table for a finite continued fraction
that approximates e =: ^ 1 .
--Kip Murray
The table below summarizes a finite continued fraction which begins
1
1 + -
1r12
1r2 +
1r60
1 + -
11 matches
Mail list logo