Some type of "normalization" process can probably be used successfully for a while, but more seriously: I think we need to track dependencies; as in, the full supply chain.
Then, perhaps, we try to figure out how to engineer solutions that will guarantee certain certain conditions will be met (e.g. Enough food for every person in an area will make it to that area for everyone there to live satisfactorily).
As the dependencies are mapped out and tracked, it will become obvious which tasks are necessary to fulfill the current economic objectives. Then, it's simply a matter of performing the tasks.
