Forcing one module per physical node in a multilayer network

[nvogtvincent]

I have a dense, directed, weighted, temporal network which I am representing as a multilayer network in Infomap. Each layer is the state of the network at a point in time, but variability is stochastic so the ordering does not matter. Nodes are persistent (i.e. the number of state nodes is equal to the number of layers for all physical nodes) - only the edge weights change between layers.

I would like to partition the network into clusters that retain flow, with the key constraint that each physical node can only be allocated to a single module (which can be multi-level). In other words, I would like to find the optimal Infomap clustering of my network subject to the constraint that the clustering must be time-invariant. Is there any way that I can do this?