[R] Package nleqslv ERROR

[J C Nash]

Maybe it's simpler to point out that nonlinear EQUATIONS need n equations for n unknowns.

4 only equals 2 through the looking glass.

Alternatively, maybe the OP wants to adjust the 2 parameters to best fit the 4 relationships, in which
case nlsr package function nlfb() would be appropriate. Note that nls() in base R tends to not like
small residual problems and tends to give other warnings and errors.


JN


On 2017-08-09 05:32 PM, Mark Sharp wrote:
> Santi,
> 
> 
> In the second line of your function you have the following:
> f <- numeric(length(x))
> 
> This sets the length of this numeric vector (i.e., "f") to the length of the vector "x".
> Later, inside the function you assign to values to 4 elements of the vector "f".
> This assumes that "f" is at least 4 element in length.
> 
> However, you define "startx" to be a numeric vector of length 2 with
> startx <- c(16350, 1.33)
> which you then use as the argument to the function "fun" (bad name for a function, by the way, as it is not descriptive).
> 
> Thus, when "x" inside your function gets the value of "startx" it becomes a numeric vector of length 2, which is then used to set the length of the numeric vector "f". As soon as the function tries to assign a value to f[3], R correctly throws an informative error.
> 
> Mark
> R. Mark Sharp, Ph.D.
> msharp at TxBiomed.org
> 
> 
> 
> 
> 
>> On Aug 9, 2017, at 3:56 PM, Santi Burone <santiagoburone at gmail.com> wrote:
>>
>> Dear all,
>> I am relatively new to R and have had some difficulty in understanding an error i get when running a code to solve a system of non-linear equations, with four equations and two variables.
>>
>> This is my code:
>>
>> ALPHA <- c(-0.0985168033402, 0.1, 0.15, 0.2, 0.25, 0.3, 0.35, 0.4)
>> BETA <- c(-0.0985168033402, 0.1, 0.15, 0.2, 0.25, 0.3, 0.35, 0.4)
>> GAMMA <- c(0.3940672148378, 0.1, 0.15, 0.2, 0.25, 0.3, 0.35, 0.4)
>> MEA <- 30000
>> MIA <- 10000
>> MAA <- 50000
>> DEA <- 0.385
>> fun <- function(x)  {
>>   f <- numeric(length(x))
>>   f[1] <-   1*x[1] - ((((MEA*((0.5*((MIA-MEA)^2))^(-BETA[1+j]))*((0.5*((MAA-MEA)^2))^(-ALPHA[1+i]))*((DEA)^(-GAMMA[1+k]))))*((12^0.5)^(-GAMMA[1+k])))/(((0.5)^(-BETA[1+j]-ALPHA[1+i]))*(((1-1*x[2])/(2))^(-BETA[1+j]))*(((-1+1*x[2])/(2))^(-ALPHA[1+i]))*(2*(1*x[2]-1)^(-GAMMA[1+k]))))^(1/(-1-BETA[1+j]-ALPHA[1+i]-1.5*GAMMA[1+k]))
>>   f[2] <-   1*x[1] - (((MAA*((0.5*((MIA+MAA)^(2)))^(-BETA[1+j]))*((DEA)^(-GAMMA[1+k])))*((12^0.5)^(-GAMMA[1+k])))/((1*x[2])*(2^(-BETA[1+j]-GAMMA[1+k]))*((1*x[2]-1)^(-GAMMA[1+k]))*((((1-1*x[2])^2)/2)^(-BETA[1+j]))))^(1/(1-2*BETA[1+j]-1.5*(GAMMA[1+k])))
>>   f[3] <-   1*x[1] - (((MIA*((0.5*((MIA+MAA)^(2)))^(-ALPHA[1+i]))*((DEA)^(-GAMMA[1+k])))*((12^0.5)^(-GAMMA[1+k])))/((2-1*x[2])*(2^(-ALPHA[1+i]-GAMMA[1+k]))*((1*x[2]-1)^(-GAMMA[1+k]))*((((-1+1*x[2])^2)/2)^(-ALPHA[1+i]))))^(1/(1-2*ALPHA[1+i]-1.5*(GAMMA[1+k])))
>>   f[4] <- 1*x[2] > 1
>>   f
>>
>> }
>>
>> Result <- matrix(0,nrow=9*9*9,ncol=6)
>> startx<-c(16350, 1.33)
>> indx <- 1
>> for (i in 1:9) {
>>   for (j in 1:9) {
>>     for (k in 1:9) {
>>       f.startx <- fun(startx)
>>       if(anyNA(f.startx)) {
>>         Result[indx,1:3] <- NA
>>       } else {
>>         z <- nleqslv(startx,fun)
>>         Result[indx,1:3] <- c(z$termcd,z$x)
>>       }
>>       Result[indx,4:6] <- c(i,j,k)
>>       indx <- indx+1
>>     }
>>   }
>> }
>>
>>
>> The error i get when solving for specific values of ALPHA, BETA and GAMMA, not using the loop is:
>>
>> Error in nleqslv(xstart, fun) : Length of fn result <> length of x!
>>
>> I had already solved this problem useing ALPHA, BETA and GAMMA as X[1], X[2] and X[3] and being X[1] and X[2] of this system  given values, in that case for the first values of alpha beta and gamma given here the solution was (16350, 1.33).
>> I dont understand what’s the error i get here, as i know the system as a unique solution. "What’s the meaning of Length of fn result <> length of x!"?
>>
>> Thanks in advance!
>> Santiago.
>> [[alternative HTML version deleted]]
>>
>> ______________________________________________
>> R-help at r-project.org mailing list -- To UNSUBSCRIBE and more, see
>> 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.
> 
> CONFIDENTIALITY NOTICE: This e-mail and any files and/or attachments transmitted, may contain privileged and confidential information and is intended solely for the exclusive use of the individual or entity to whom it is addressed. If you are not the intended recipient, you are hereby notified that any review, dissemination, distribution or copying of this e-mail and/or attachments is strictly prohibited. If you have received this e-mail in error, please immediately notify the sender stating that this transmission was misdirected; return the e-mail to sender; destroy all paper copies and delete all electronic copies from your system without disclosing its contents.
> ______________________________________________
> R-help at r-project.org mailing list -- To UNSUBSCRIBE and more, see
> 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.
>