Jorge Ivan Velez
2008-Jun-03 21:09 UTC
[R] How to solve a non-linear system of equations using R
Dear R-list members,
I've had a hard time trying to solve a non-linear system (nls) of equations
which structure for the equation i, i=1,...,4, is as follows:
f_i(d_1,d_2,d_3,d_4)-k_i(l,m,s) = 0 (1)
In the expression above, both f_i and k_i are known functions and l, m and s
are known constants. I would like to estimate the vector d=(d_1,d_2,d_3,d_4)
which is solution of (1). Functions in R to estimate f_i-k_i are at the end
of this message.
Any help/suggestions/comments would be greatly appreciated.
Thanks in advance,
Jorge
# ------------------------------
# Constants
# ------------------------------
l=1
m=0.4795
s=0.4795
# ------------------------------
# Functions to estimate f_i-k_i
# ------------------------------
f1=function(d){
d1=d[1]
d2=d[2]
d3=d[3]
d4=d[4]
res=2*d1+2*sqrt(2)*d1*d2+2*sqrt(3)*d2*d3+4*d3*d4-l*m*(1+d1^2+d2^2+d3^2+d4^2)
res
}
f2=function(d){
d1=d[1]
d2=d[2]
d3=d[3]
d4=d[4]
res=2*sqrt(2)*d2+2*d1^2+2*sqrt(6)*d1*d3+4*d2^2+4*sqrt(3)*d2*d4+6*d3^2+8*d4^2-l*(m^2+m^3*s^(-1))*(1+d1^2+d2^2+d3^2+d4^2)
res
}
f3=function(d){
d1=d[1]
d2=d[2]
d3=d[3]
d4=d[4]
res=6*d1+12*sqrt(2)*d1*d2+18*sqrt(3)*d2*d3+48*d3*d4+2*sqrt(6)*d3+4*sqrt(6)*d1*d4-l*(m^3+3*m^4*s^(-1)+3*m^6*s^(-2))*(1+d1^2+d2^2+d3^2+d4^2)
res
}
f4=function(d){
d1=d[1]
d2=d[2]
d3=d[3]
d4=d[4]
res=12*sqrt(2)*d2+12*d1^2+36*d2^2+72*d3^2+120*d4^2+20*sqrt(6)*d1*d3+56*sqrt(3)*d2*d4+4*sqrt(6)*d4-l*((m^4+6*m^6*s^(-1)+15*m^6*s^(-2)+15*m^7*s^(-3))-3*l^(2)*(m^2+m^3*s^(-1))^2)*(1+d1^2+d2^2+d3^2+d4^2)
res
}
[[alternative HTML version deleted]]
ctu at bigred.unl.edu
2008-Jun-03 21:22 UTC
[R] How to solve a non-linear system of equations using R
Hi Jorge, Have you tried to use "systemfit" package. In this package, this is a function call " nlsystemfit ". This might help. Chunhao Quoting Jorge Ivan Velez <jorgeivanvelez at gmail.com>:> Dear R-list members, > > I've had a hard time trying to solve a non-linear system (nls) of equations > which structure for the equation i, i=1,...,4, is as follows: > > > f_i(d_1,d_2,d_3,d_4)-k_i(l,m,s) = 0 (1) > > > In the expression above, both f_i and k_i are known functions and l, m and s > are known constants. I would like to estimate the vector d=(d_1,d_2,d_3,d_4) > which is solution of (1). Functions in R to estimate f_i-k_i are at the end > of this message. > > Any help/suggestions/comments would be greatly appreciated. > > Thanks in advance, > > Jorge > > > # ------------------------------ > # Constants > # ------------------------------ > > l=1 > m=0.4795 > s=0.4795 > > # ------------------------------ > # Functions to estimate f_i-k_i > # ------------------------------ > > f1=function(d){ > d1=d[1] > d2=d[2] > d3=d[3] > d4=d[4] > res=2*d1+2*sqrt(2)*d1*d2+2*sqrt(3)*d2*d3+4*d3*d4-l*m*(1+d1^2+d2^2+d3^2+d4^2) > res > } > > f2=function(d){ > d1=d[1] > d2=d[2] > d3=d[3] > d4=d[4] > res=2*sqrt(2)*d2+2*d1^2+2*sqrt(6)*d1*d3+4*d2^2+4*sqrt(3)*d2*d4+6*d3^2+8*d4^2-l*(m^2+m^3*s^(-1))*(1+d1^2+d2^2+d3^2+d4^2) > res > } > > f3=function(d){ > d1=d[1] > d2=d[2] > d3=d[3] > d4=d[4] > res=6*d1+12*sqrt(2)*d1*d2+18*sqrt(3)*d2*d3+48*d3*d4+2*sqrt(6)*d3+4*sqrt(6)*d1*d4-l*(m^3+3*m^4*s^(-1)+3*m^6*s^(-2))*(1+d1^2+d2^2+d3^2+d4^2) > res > } > > > f4=function(d){ > d1=d[1] > d2=d[2] > d3=d[3] > d4=d[4] > res=12*sqrt(2)*d2+12*d1^2+36*d2^2+72*d3^2+120*d4^2+20*sqrt(6)*d1*d3+56*sqrt(3)*d2*d4+4*sqrt(6)*d4-l*((m^4+6*m^6*s^(-1)+15*m^6*s^(-2)+15*m^7*s^(-3))-3*l^(2)*(m^2+m^3*s^(-1))^2)*(1+d1^2+d2^2+d3^2+d4^2) > res > } > > [[alternative HTML version deleted]] > > ______________________________________________ > R-help at 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. >
Moshe Olshansky
2008-Jun-04 00:00 UTC
[R] How to solve a non-linear system of equations using R
Since k_i(l,m,s) are known constants, you actually have a system of four non-linear equations with 4 unknowns. One possibility is to use optim (check ?optim). Another one is to use the very recently released package - look at https://stat.ethz.ch/pipermail/r-help/attachments/20080423/da0b7f6c/attachment.pl --- On Wed, 4/6/08, Jorge Ivan Velez <jorgeivanvelez at gmail.com> wrote:> From: Jorge Ivan Velez <jorgeivanvelez at gmail.com> > Subject: [R] How to solve a non-linear system of equations using R > To: "R mailing list" <r-help at r-project.org> > Received: Wednesday, 4 June, 2008, 7:09 AM > Dear R-list members, > > I've had a hard time trying to solve a non-linear > system (nls) of equations > which structure for the equation i, i=1,...,4, is as > follows: > > > f_i(d_1,d_2,d_3,d_4)-k_i(l,m,s) = 0 (1) > > > In the expression above, both f_i and k_i are known > functions and l, m and s > are known constants. I would like to estimate the vector > d=(d_1,d_2,d_3,d_4) > which is solution of (1). Functions in R to estimate > f_i-k_i are at the end > of this message. > > Any help/suggestions/comments would be greatly appreciated. > > Thanks in advance, > > Jorge > > > # ------------------------------ > # Constants > # ------------------------------ > > l=1 > m=0.4795 > s=0.4795 > > # ------------------------------ > # Functions to estimate f_i-k_i > # ------------------------------ > > f1=function(d){ > d1=d[1] > d2=d[2] > d3=d[3] > d4=d[4] > res=2*d1+2*sqrt(2)*d1*d2+2*sqrt(3)*d2*d3+4*d3*d4-l*m*(1+d1^2+d2^2+d3^2+d4^2) > res > } > > f2=function(d){ > d1=d[1] > d2=d[2] > d3=d[3] > d4=d[4] > res=2*sqrt(2)*d2+2*d1^2+2*sqrt(6)*d1*d3+4*d2^2+4*sqrt(3)*d2*d4+6*d3^2+8*d4^2-l*(m^2+m^3*s^(-1))*(1+d1^2+d2^2+d3^2+d4^2) > res > } > > f3=function(d){ > d1=d[1] > d2=d[2] > d3=d[3] > d4=d[4] > res=6*d1+12*sqrt(2)*d1*d2+18*sqrt(3)*d2*d3+48*d3*d4+2*sqrt(6)*d3+4*sqrt(6)*d1*d4-l*(m^3+3*m^4*s^(-1)+3*m^6*s^(-2))*(1+d1^2+d2^2+d3^2+d4^2) > res > } > > > f4=function(d){ > d1=d[1] > d2=d[2] > d3=d[3] > d4=d[4] > res=12*sqrt(2)*d2+12*d1^2+36*d2^2+72*d3^2+120*d4^2+20*sqrt(6)*d1*d3+56*sqrt(3)*d2*d4+4*sqrt(6)*d4-l*((m^4+6*m^6*s^(-1)+15*m^6*s^(-2)+15*m^7*s^(-3))-3*l^(2)*(m^2+m^3*s^(-1))^2)*(1+d1^2+d2^2+d3^2+d4^2) > res > } > > [[alternative HTML version deleted]] > > ______________________________________________ > R-help at 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.
Ravi Varadhan
2008-Jun-04 18:55 UTC
[R] How to solve a non-linear system of equations using R
Jorge,
You can use the package "BB" to try and solve this problem.
I have re-written your functions a little bit.
# ------------------------------
# Constants
# ------------------------------
l=1
m=0.4795
s=0.4795
# ------------------------------
# Functions to estimate f_i-k_i
# ------------------------------
myfn <- function(d){
d1 <- d[1]
d2 <- d[2]
d3 <- d[3]
d4 <- d[4]
res <- rep(NA, 4)
res[1] <-
2*d1+2*sqrt(2)*d1*d2+2*sqrt(3)*d2*d3+4*d3*d4-l*m*(1+d1^2+d2^2+d3^2+d4^2)
res[2] <-
2*sqrt(2)*d2+2*d1^2+2*sqrt(6)*d1*d3+4*d2^2+4*sqrt(3)*d2*d4+6*d3^2+8*d4^2-l*(
m^2+m^3*s^(-1))*(1+d1^2+d2^2+d3^2+d4^2)
res[3] <-
6*d1+12*sqrt(2)*d1*d2+18*sqrt(3)*d2*d3+48*d3*d4+2*sqrt(6)*d3+4*sqrt(6)*d1*d4
-l*(m^3+3*m^4*s^(-1)+3*m^6*s^(-2))*(1+d1^2+d2^2+d3^2+d4^2)
res[4] <-
12*sqrt(2)*d2+12*d1^2+36*d2^2+72*d3^2+120*d4^2+20*sqrt(6)*d1*d3+56*sqrt(3)*d
2*d4+4*sqrt(6)*d4-l*((m^4+6*m^6*s^(-1)+15*m^6*s^(-2)+15*m^7*s^(-3))-3*l^(2)*
(m^2+m^3*s^(-1))^2)*(1+d1^2+d2^2+d3^2+d4^2)
res
}
myfn.opt <- function(d){
# re-writing "myfn" to be used for minimization
d1 <- d[1]
d2 <- d[2]
d3 <- d[3]
d4 <- d[4]
res <- rep(NA, 4)
res[1] <-
2*d1+2*sqrt(2)*d1*d2+2*sqrt(3)*d2*d3+4*d3*d4-l*m*(1+d1^2+d2^2+d3^2+d4^2)
res[2] <-
2*sqrt(2)*d2+2*d1^2+2*sqrt(6)*d1*d3+4*d2^2+4*sqrt(3)*d2*d4+6*d3^2+8*d4^2-l*(
m^2+m^3*s^(-1))*(1+d1^2+d2^2+d3^2+d4^2)
res[3] <-
6*d1+12*sqrt(2)*d1*d2+18*sqrt(3)*d2*d3+48*d3*d4+2*sqrt(6)*d3+4*sqrt(6)*d1*d4
-l*(m^3+3*m^4*s^(-1)+3*m^6*s^(-2))*(1+d1^2+d2^2+d3^2+d4^2)
res[4] <-
12*sqrt(2)*d2+12*d1^2+36*d2^2+72*d3^2+120*d4^2+20*sqrt(6)*d1*d3+56*sqrt(3)*d
2*d4+4*sqrt(6)*d4-l*((m^4+6*m^6*s^(-1)+15*m^6*s^(-2)+15*m^7*s^(-3))-3*l^(2)*
(m^2+m^3*s^(-1))^2)*(1+d1^2+d2^2+d3^2+d4^2)
sum(res^2)
}
library(BB)
p0 <- runif(4, -1,0)
ans1 <- dfsane(par=p0, fn=myfn)
ans2 <- spg(par=p0, fn=myfn.opt)
ans1
ans2
Note that the above does not produce a redual of zero, so the system can't
be solved exactly. I tried a large number of random starting values without
improving upon the solution provided by "spg". So, you may want to
check
your system for its correctness.
Hope this helps,
Ravi.
-----Original Message-----
From: r-help-bounces at r-project.org [mailto:r-help-bounces at r-project.org]
On
Behalf Of Jorge Ivan Velez
Sent: Tuesday, June 03, 2008 5:09 PM
To: R mailing list
Subject: [R] How to solve a non-linear system of equations using R
Dear R-list members,
I've had a hard time trying to solve a non-linear system (nls) of equations
which structure for the equation i, i=1,...,4, is as follows:
f_i(d_1,d_2,d_3,d_4)-k_i(l,m,s) = 0 (1)
In the expression above, both f_i and k_i are known functions and l, m and s
are known constants. I would like to estimate the vector d=(d_1,d_2,d_3,d_4)
which is solution of (1). Functions in R to estimate f_i-k_i are at the end
of this message.
Any help/suggestions/comments would be greatly appreciated.
Thanks in advance,
Jorge
# ------------------------------
# Constants
# ------------------------------
l=1
m=0.4795
s=0.4795
# ------------------------------
# Functions to estimate f_i-k_i
# ------------------------------
f1=function(d){
d1=d[1]
d2=d[2]
d3=d[3]
d4=d[4]
res=2*d1+2*sqrt(2)*d1*d2+2*sqrt(3)*d2*d3+4*d3*d4-l*m*(1+d1^2+d2^2+d3^2+d4^2)
res
}
f2=function(d){
d1=d[1]
d2=d[2]
d3=d[3]
d4=d[4]
res=2*sqrt(2)*d2+2*d1^2+2*sqrt(6)*d1*d3+4*d2^2+4*sqrt(3)*d2*d4+6*d3^2+8*d4^2
-l*(m^2+m^3*s^(-1))*(1+d1^2+d2^2+d3^2+d4^2)
res
}
f3=function(d){
d1=d[1]
d2=d[2]
d3=d[3]
d4=d[4]
res=6*d1+12*sqrt(2)*d1*d2+18*sqrt(3)*d2*d3+48*d3*d4+2*sqrt(6)*d3+4*sqrt(6)*d
1*d4-l*(m^3+3*m^4*s^(-1)+3*m^6*s^(-2))*(1+d1^2+d2^2+d3^2+d4^2)
res
}
f4=function(d){
d1=d[1]
d2=d[2]
d3=d[3]
d4=d[4]
res=12*sqrt(2)*d2+12*d1^2+36*d2^2+72*d3^2+120*d4^2+20*sqrt(6)*d1*d3+56*sqrt(
3)*d2*d4+4*sqrt(6)*d4-l*((m^4+6*m^6*s^(-1)+15*m^6*s^(-2)+15*m^7*s^(-3))-3*l^
(2)*(m^2+m^3*s^(-1))^2)*(1+d1^2+d2^2+d3^2+d4^2)
res
}
[[alternative HTML version deleted]]
______________________________________________
R-help at 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.