Hi
Do you mean 4x4 matrix or you mean a 4-dimensional matrix. I think you mean the latter.
Here is how if I understood correctly:
#data file, change values below as you want
param nK:= 3; #number of commodoties
param nN:= 4; #number of nodes
param nT:= 8; #number of time instances/values
#model file
param nK>=0;
param nN>=0;
param nT>=0;
set K:= 1..nK;
set N:= 1..nN;
set T:= 1..nT;
param cost{i in N, j in N, k in K, t in T} >=0;
var F{i in N, j in N, k in K, t in T} >=0 binary;
var FCR>=0;
#objective function here
#constraints here
since the equation for the objective is fractional you need to manipulate it. I had a similar issue so I cross-multiplied so you will have
FCR x sum{F_ijkt} = sum{cost_ijkt}x
F_ijkt
The left hand side of the equation is then quadratic where you multiply two variables (I assumed the F_ijkt is binary indicating if the i,j path is used, this also makes the model easier). Look for the linearization of a product of a binary variable times a coninuous variable and then add your remaining constraints and you should be able to complete it.
All the best.