data processing in Cadence

Hi,
I have a question and I'll try to break it down:
I have RC circuit, connected from the input side with a VerilogA model and let us called it input model, this input model read data from .dat file and sum it with another signals (I'll explain later from where does this signal come from).
Let us call the previous circuit "cell number one".
I should implement this cell 256 times as a 2 D 16 x 16 matrix, since each cell represents a pixel in an image .
Each cell receives signal from its neighbors (that is the signal I mentioned before)
and because the position of the cell will play a role from which neighbor it will receive the signal, in this way I cannot do just save as the file, I must go and connect the circuit to only its neighbors, it is a long procedures, I'm asking is there a way easier than that?

The other thing is the middle cell in this 16 x 16 matrix , this middle cell should give +1 V, if not I should measure the actual output of that cell (middle) and calculate the square error and return back to the input model to change some parameters (using any training method of the cellular neural networks), my question here is the error measuring and going back to the input model should be implemented in a loop-like way in the programming, what I did, I increased the final time of the transient analysis to give the simulator a chance to have as many loops as possible, is there another way to do this?

thank you