[R] Solve system of non linear equations using nasted loops

[David Winsemius]

> On May 10, 2017, at 11:20 AM, Santiago Burone <santiagoburone at icloud.com> wrote:
> 
> Hello, 
> 
> I'm new at R and I would like to use it in order to solve a system of non linear equations. I have the code that works but im not able to save the results. 
> 
> 
> 
> My system has three equations and i would like to solve this using three nested loops, becouse i need solutions for all the values combinations.
> 
> 
> 
> 
> 
> The code im using is this:
> 
> 
> 
> library(nleqslv)
> #x3 es gamma
> #x2 es alpha
> #x1 es beta
> MA <- c(50000, 43600, 40000, 38800, 37600, 34400, 31600, 27200, 24400, 20000)
> MI <- c(10000, 21800, 20000, 19400, 18800, 17200, 15800, 13600, 12200, 10000)
> ME <- c(30000, 32700, 30000, 29100, 28200, 25800, 23700, 20400, 18300, 15000)
> DE <- c(0.384900179, 0.19245009, 0.19245009, 0.19245009, 0.19245009, 0.19245009, 0.19245009, 0.19245009, 0.19245009, 0.19245009)
> for (i in 1:9) {
> for (j in 1:9){
> for (k in 1:9){

I got an error on the first pass through the loop.


> df.names <- paste("SociedadB",1:10,sep="")

# seems kind of excessive to repeat that process 9*9*9 times
# Maybe this should be done outside the loop?

> 
> fun <- function(x) { 
> f <- numeric(length(x)) # read as:
> f[1] <- 1*x[3] - (log(ME[1+i]/ME[1])+(1*x[1]+1*x[2])*(2)*(log((MA[1]-ME[1])/(MA[i+1]-ME[i+1]))))/(log(0.19245009/0.384900179))
> f[2] <- 1*x[3] - (log(MI[1+j]/MI[1])-(1*x[2])*(2)*(log((MA[1+j]-MI[1+j])/(MA[1]-MI[1]))))/(log(0.19245009/0.384900179)) 
> f[3] <- 1*x[3] - (log(MA[1+k]/MA[1])-(1*x[1])*(2)*(log((MI[1+k]-MA[1+k])/(MI[1]-MA[1]))))/(log(0.19245009/0.384900179)) 
> f 
> } 

# So you need 9*9*9 different functions? 


> startx <- c(1,1,1) 
> answers<-as.data.frame(nleqslv(startx,fun)) 

> d.frame <- (answers$x)
> assign(df.names[k+20], d.frame)

# I don't think that df.names[k+20] would be anything other than NA since it only has 10 elements

> despejes<- data.frame(SociedadB2, SociedadB3, SociedadB4, SociedadB5, SociedadB6, SociedadB7, SociedadB8, SociedadB9, SociedadB10)
> 
> }
> }
> }
> 
> despejes<- data.frame(SociedadB2, SociedadB3, SociedadB4, SociedadB5, SociedadB6, SociedadB7, SociedadB8, SociedadB9, SociedadB10)
> 

The error message is:

Error in data.frame(SociedadB2, SociedadB3, SociedadB4, SociedadB5, SociedadB6,  : 
  object 'SociedadB2' not found

That's because column names are not first class data-objects in R.


> 
> 
> If im not wrong i should have 1000 different solutions but the data frame only save 10 results. I tried copying the last part of the code between the last two brackets (changing the name of the data frame despejes for despejes 1, 2 and 3 but i only get three data frames with exactly the same results. I dont know how could i save all the results of every combination. I mean, i need to solve for i=1, j=1 and the 9 diferent values of k and save those results, then i=1, j=2 and the 9 values of k..
> 

If you need to save separate results for each combination of i,j,k , then you need at a minimum an array (possibly with a further 4th dimension if you are storing numerical vectors)  or possibly a list if these items are more complicated than an atomic vector.
> 
> 
> I know this might be a easy question but i have not been able to find a solution in the last days. 
> 
> 
> 
> Thanks in advance if you can help me.
> 
> 
> Santiago. 
> 	[[alternative HTML version deleted]]

Rhelp is a plain text list although you don't seem to have run into trouble on that account. You should however determine how your mail client can be setup to just send pain text to this mailing list.


> 
> ______________________________________________
> 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.

David Winsemius
Alameda, CA, USA