Thank you for the generous help, it is greatly appreciated as I’ve been scratching my head thinking about this one for a while (and am new to Julia and JuMP).
No problem.
I have added a dimension to the y and u matrices to accommodate the multiple networks.
it essentially repeats the tsp for each source on the same static distance matrix for all the nodes
This seems like the right approach for now. See also:
Is there a way to modify the constraints such that the TSP is run iteratively
There are other formulations, see Traveling Salesperson Problem · JuMP, but I’d caution against jumping (if you will) straight to that. I’d get the MTZ formulation working for small instances, and then go from there once you have a handle on the problem.