Thank you for your answer.
I posted this problem as it is because I am benchmarking multiple solvers
over this particular problem.
I will enquire LowRankQP, kernlab and quadprog packages as you suggested.
Thank you
Georges
Spencer Graves wrote:
>
> I believe that 'optim' will not accept equality constraints.
>
> However, you do not need the generality of 'optim' to "minimize a
> quadratic function with boundary conditions and one equality
> condition". This type of problem is called "quadratic programming",
> and RSiteSearch("quadratic programming", "fun") just returned 29 hits
> for me. The first 3 cite functions in the "LowRankQP", "kernlab", and
> "quadprog" packages. I don't know if any of these will solve your
> problem, but I suspect that at least one might. If not, can you recast
> the problem to remove the equality constraint?
>
> If the above does not work for you, I suggest you try a much
> simpler version, e.g., with 'mat' = a 3 x 3 array with one inequality
> and one equality, as suggested in the famous book by Polya on "How to
> Solve It" (http://en.wikipedia.org/wiki/How_to_Solve_It). This has the
> added advantage of giving you a simpler example to send to this list if
> you can't make it work. You are to be commended for providing a
> self-contained example. Unfortunately, your example is so large that it
> is slightly intimidating. A simpler example might elicit more (and more
> useful) replies -- if it doesn't lead you to the solution, as Polya
> suggested that it might.
>
> Hope this helps,
> Spencer
>
> lhaba wrote:
>> Hi,
>> i need to minimize a quadratic function with boundary condidtions and one
>> equality condition.
>> In order to do that i converted the equality constraint into 2 inequality
>> constaints and passed everything cia constrOptim, as the manual said:
>> everything included in the ... will be passed to Optim that will pass it
>> back to fn in case it does not need it.
>>
>> My code is the following:
>>
>> mat <- array( c(0.0001088799073581, 0.0000136029502036,
>> 0.0000060430384243,
>> 0.0000847097879033, 0.0000115053365822, 0.0000216245975292,
>> 0.0000483253391811, 0.0000787580901352, 0.0000186474817658,
>> 0.0000312260571354, 0.0000217594093734, 0.0000536298150897,
>> 0.0000166202592455, 0.0000451975061637, -0.0000120364862228,
>> 0.0000497117714376,
>> 0.0000136029502036, 0.0001537319301276, 0.0000226518408080,
>> 0.0000591480002102, 0.0000797128619950, 0.0000091332643423,
>> 0.0000693354260457, 0.0000825217915015, 0.0000229122227269,
>> 0.0000297662414650, 0.0000334443258658, 0.0000273254534933,
>> 0.0000202062301763, 0.0000026260702295, 0.0000558975248740,
>> 0.0000953647537111,
>> 0.0000060430384243, 0.0000226518408080, 0.0005971325756834,
>> -0.0000762583321100, -0.0000246005202071, -0.0000300982253054,
>> 0.0000299178429478, 0.0000135672602503, 0.0001735431064391,
>> -0.0000133347388414, 0.0001387582890571, 0.0000964898243724,
>> -0.0000149571346672, 0.0000104437939143, 0.0001246900353191,
>> -0.0000171884354549,
>> 0.0000847097879033, 0.0000591480002102, -0.0000762583321100,
>> 0.0004968467836203, 0.0002303499425964, 0.0000992731601466,
>> 0.0002685466918035, 0.0002580180069951, 0.0000725833959653,
>> 0.0000525639940758, 0.0001785049461665, 0.0001781339191317,
>> 0.0000597631329497, 0.0000201160486244, 0.0002582267884874,
>> 0.0002473268250781,
>> 0.0000115053365822, 0.0000797128619950, -0.0000246005202071,
>> 0.0002303499425964, 0.0002945009393242, -0.0000426583313588,
>> 0.0002067711081561, 0.0002695894499975, 0.0001312519434236,
>> -0.0000079156628396, 0.0001423655606105, 0.0000044733483182,
>> 0.0000303832556655, 0.0000577624190434, 0.0001193435284164,
>> 0.0002422477575812,
>> 0.0000216245975292, 0.0000091332643423, -0.0000300982253054,
>> 0.0000992731601466, -0.0000426583313588, 0.0001641146317929,
>> 0.0000311621614693, -0.0000147821020927, -0.0000767394607354,
>> 0.0000619936562782, -0.0000306228761064, 0.0001495752154579,
>> 0.0000389317919640, -0.0000714551280935, -0.0000564616194935,
>> 0.0000384367900903,
>> 0.0000483253391811, 0.0000693354260457, 0.0000299178429478,
>> 0.0002685466918035, 0.0002067711081561, 0.0000311621614693,
>> 0.0003176493360736, 0.0002575792630182, 0.0001371966488704,
>> 0.0000436833885846, 0.0001442516276721, 0.0001075447728937,
>> 0.0000371155448252, 0.0000475873370276, 0.0002162409964174,
>> 0.0002870514043081,
>> 0.0000787580901352, 0.0000825217915015, 0.0000135672602503,
>> 0.0002580180069951, 0.0002695894499975, -0.0000147821020927,
>> 0.0002575792630182, 0.0006217963583393, 0.0002368375072233,
>> 0.0000078625467985, 0.0002054774387807, -0.0000066572248626,
>> 0.0000485854317294, 0.0002802199677114, 0.0001676465030622,
>> 0.0003028775764026,
>> 0.0000186474817658, 0.0000229122227269, 0.0001735431064391,
>> 0.0000725833959653, 0.0001312519434236, -0.0000767394607354,
>> 0.0001371966488704, 0.0002368375072233, 0.0004475645060339,
>> -0.0000030389778729, 0.0001706183643212, -0.0000017789896670,
>> 0.0000722657436668, 0.0001664088523103, 0.0001220193496918,
>> 0.0001641280878243,
>> 0.0000312260571354, 0.0000297662414650, -0.0000133347388414,
>> 0.0000525639940758, -0.0000079156628396, 0.0000619936562782,
>> 0.0000436833885846, 0.0000078625467985, -0.0000030389778729,
>> 0.0000822356406019, -0.0000226786278360, 0.0000752056105897,
>> 0.0000399801889185, -0.0000441549693477, 0.0000047887593401,
>> 0.0000352165734549,
>> 0.0000217594093734, 0.0000334443258658, 0.0001387582890571,
>> 0.0001785049461665, 0.0001423655606105, -0.0000306228761064,
>> 0.0001442516276721, 0.0002054774387807, 0.0001706183643212,
>> -0.0000226786278360, 0.0004304869804941, 0.0001566983136020,
>> 0.0000332770114864, 0.0000012432094922, 0.0002491186667930,
>> 0.0001285479414542,
>> 0.0000536298150897, 0.0000273254534933, 0.0000964898243724,
>> 0.0001781339191317, 0.0000044733483182, 0.0001495752154579,
>> 0.0001075447728937, -0.0000066572248626, -0.0000017789896670,
>> 0.0000752056105897, 0.0001566983136020, 0.0005292416268831,
>> 0.0000893358436932, -0.0001009559617338, 0.0000888461032129,
>> 0.0000714719761291,
>> 0.0000166202592455, 0.0000202062301763, -0.0000149571346672,
>> 0.0000597631329497, 0.0000303832556655, 0.0000389317919640,
>> 0.0000371155448252, 0.0000485854317294, 0.0000722657436668,
>> 0.0000399801889185, 0.0000332770114864, 0.0000893358436932,
>> 0.0001844874143317, 0.0000549019705905, -0.0000117658984941,
>> 0.0000394986211508,
>> 0.0000451975061637, 0.0000026260702295, 0.0000104437939143,
>> 0.0000201160486244, 0.0000577624190434, -0.0000714551280935,
>> 0.0000475873370276, 0.0002802199677114, 0.0001664088523103,
>> -0.0000441549693477, 0.0000012432094922, -0.0001009559617338,
>> 0.0000549019705905, 0.0006686123611712, -0.0001115788528761,
>> 0.0000151312169512,
>> -0.0000120364862228, 0.0000558975248740, 0.0001246900353191,
>> 0.0002582267884874, 0.0001193435284164, -0.0000564616194935,
>> 0.0002162409964174, 0.0001676465030622, 0.0001220193496918,
>> 0.0000047887593401, 0.0002491186667930, 0.0000888461032129,
>> -0.0000117658984941, -0.0001115788528761, 0.0014937840813054,
>> 0.0001299625832782,
>> 0.0000497117714376, 0.0000953647537111, -0.0000171884354549,
>> 0.0002473268250781, 0.0002422477575812, 0.0000384367900903,
>> 0.0002870514043081, 0.0003028775764026, 0.0001641280878243,
>> 0.0000352165734549, 0.0001285479414542, 0.0000714719761291,
>> 0.0000394986211508, 0.0000151312169512, 0.0001299625832782,
>> 0.0004355778955394), c(16,16))
>>
>> covar <- function(x) return (t(x) %*%mat %*% (x))
>> covargr <- function(x) return ( 2*mat %*% (x))
>>
>> upper1 = c(1, 0, 0, 0, 0, 0, 0,0 ,0 , 0, 0, 0, 0, 0, 0, 0)
>> lower1 = c(1, -0.25, -0.25, -0.25, -0.25, -0.25, -0.25, -0.25, -0.25,
>> -0.25,
>> -0.25, -0.25, -0.25, -0.25, -0.25, -0.25)
>> lower[1] = 1
>> init = upper1- 1/15.0
>> init[1] = 1
>>
>> optim( init, covar, covargr, method = "L-BFGS-B", lower = lower1, upper
>> =
>> upper1)
>> u1 <- c(0, 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1, 0,
>> -1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1)
>> u2<- array(u1, c(16, 2))
>> u2[,2] <- -u2[,2]
>>
>> u1<- t(u1)
>> c1<- array(c(-1, -1),c(2,1))
>> constrOptim (init, covar, covargr, t(u2), c1, mu = 1e-054, method =
>> "L-BFGS-B", lower = lower1, upper = upper1, outer.iterations = 100,
>> outer.eps = 1e-05)
>>
>>
>> The problem I face is an error message telling me that the method
>> "L-BFGS-B"
>> needs finite values of fn...
>> while everything works well in the optim case without the inequalities.
>>
>> Does anybody have any clue about what might have gone wrong?
>>
>> THank you for your help
>>
>>
>> Georges
>>
>>
>
> ______________________________________________
> R-help@r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide
> http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.
>
>
--
View this message in context:
http://www.nabble.com/constrOptim-with-method-%3D-%22L-BFGS-B%22-tp17916460p17961273.html
Sent from the R help mailing list archive at Nabble.com.
______________________________________________
R-help@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.
