How to calculate the resistance of a resistors in a grid? I've found a lot of example of a grid where you want to total resistance of a grid. My problem is different. Let's say you got a grid of resistance ($3\times 3$) but you only have the total resistance between each combination of line, is it possible find the value of the resistance?
Let's take this grid for instance : 

As you can see, the resistances R1 to R9 got different value (10 Ohms to 90 Ohms). Let's say you can only have the total resistance between each point (A to F), is it possible, mathematically speaking, to find the Resistance R1 to R9?
I've used a simulator to get the total resistance between each point. Here it is :

So, in conclusion, the question is : "Is it possible, with the value of that table, to find back the individual value of R1 to R9?
 A: By connecting A with B and E with F you can reduce the problem to a 2x2 matrix. Same for all other combinations of two (or n-1 lines in general). I would then call the resistors $R_{11}$ to $R_{22}$. If you short A with B and E with F, you combine resistors in such a way that $R_{22}=R_9$. Shorting out a horizontal line with a vertical line in the simplified circuit removes one of the 4 resistors from the problem. The resulting circuit is the parallel combination of one resistor (pick e.g. $R_{22}$) and the sum of two others ($R_{11}$ and $R_{12}$) which also form a voltage divider for which the central voltage is known. This way you have enough equations to calculate all three resistors in the simplified circuit, from which you then determine $R_{22}=R_9$. Rinse and repeat by combining other horizontal and vertical lines. 
This is not the most precise or most efficient method for a real problem of this kind, of course, especially if a lot of rows and columns exist and the shorting of most of them will lead to a really low parallel resistance in the final 2x2 problem. It's a simple way of seeing that the problem is solvable, though. A better method would probably use individual charge injection into rows while columns are being held on constant voltage (and vice versa). 
A: If you ask this question in full generality -- i.e., is it possible for any value of R1 to R9 to infer their value from the resistances between A..F -- then answer is no.
Just imagine $R_1=R_2=R_4=0$.  You will never be able to infer $R_5$, as there is a short circuit.
Regardless of that, I doubt that there is generically a unique solution, since the system of equations is non-linear.
