Some random remarks about your paper.

Your vision about hub's remuneration is a bit short-sighted, you don't consider that a hub can be tied to a custom version of wallet from which the operator can make revenue from advertizing or collected data. Gmail, Whatapps or whatever make money although having to support high infrastructure cost. If the user doesn't want to be the product, he could still choose to volontary pay for a hub, a small payment is so easy with Byteball... I don't see why this should be included in the protocol.

This is a bit weird when you say that nobody would like to be a witness since you've been trying for 1 year to become one and I reckon that anybody somehow involved with Byteball would gladly accept to be one. Maybe your arguments could be right but I'm not convinced by witnesses chosen by proof of stake, it would give power to a few rich holders from whom we know nothing and remove the possibility for users to get a witness out. In overall this solution would suffer from all the flaws of PoS (see PoS vs PoW debate).

When you want to remove the header commission and keep the DAG narrow with a hard coded function, have you an idea of the algorithm used ? it doesn't look trivial to do.