Generalization of QF/QV for continuous time?

Suppose I have 3 computers on a network, and I want to fairly allocate the computers across 5 users in real-time (so that when a user decides they really want a computer right now, they can get one, if they spend enough of their credits).

Is there a generalization of QF/QV designed for this kind of situation?

1 Like

I feel like a Harberger tax would be a better solution than QF/QV for this usecase.

i.e. each user earns x time credits per day which they use to buy time on the machine. The current user set’s the price at which they would be happy to lose access to the machine and then pays y% tax per time period to keep control of the system.
At any point, another user can take control of the system by paying the current user x tokens and then setting their own x.
If at any point the current user no longer wishes to use the system, they can stop paying their tax, which initiates a dutch auction for the to decide the new user.

One thing worth pointing out is that this is probably optimizing for allocative efficiency than fairness, but assuming all users earn time credits at the same rate (or some other “fair” distribution), then it should be “fair” as well.