I'm using lpsolveAPI in R and would like to set coefficients for specified columns and rows (coefficient for specified constraint number and decision variable number).
However, whereas I can add new column (new decision variable) or set existing column, I can't edit the column as that option will remove all previous coefficients in that column.
For example, let it be 5 constraints and 2 decision variables. Then:
lps.model <- make.lp(5, 2) #create lp model
#set coefficients for the first 3 constraint for both variables
for (i in seq(1,2)) set.column(lps.model, i, c(1,2,3), indices = c(1,2,3))
The model looks like this:
Model name:
C1 C2 C3 C4
Minimize 0 0 0 0
R1 1 1 0 0 free 0
R2 2 2 0 0 free 0
R3 3 3 0 0 free 0
R4 0 0 0 0 free 0
R5 0 0 0 0 free 0
Kind Std Std Std Std
Type Real Real Real Real
Upper Inf Inf Inf Inf
Lower 0 0 0 0
Now I want to add coefficients for 4th and 5th constraints.
for (i in seq(1,2)) set.column(lps.model, i, c(4,5), indices = c(4,5))
The code will rewrite the model since set.column functions sets all coefficients that were not listed within the function parameters to 0.
Model name:
C1 C2
Minimize 0 0
R1 0 0 free 0
R2 0 0 free 0
R3 0 0 free 0
R4 4 4 free 0
R5 5 5 free 0
Kind Std Std
Type Real Real
Upper Inf Inf
Lower 0 0
I have a big matrix of constraints and decision variables and need to run alike loops for different sets of variables. Is there any way to edit existing columns without rewriting them?
You could use set.mat for this to set values in your A matrix one at a time. See the help here.
For example:
> set.mat(lps.model, 4,5,3)
Will make the value of 4th row, 5th column to be 3, without overwriting anything else. So you can call set.mat within a double loop and change individual values.
However, it would be much more efficient if you can create whole columns at a time (preprocessing to create the list of coefficients) and then adding them to the lps.model in one shot using set.column especially since you say you have a large matrix of decision variables.