Quantum search enables square-sped up search for marked element. When there are multiple maked element, grover search provides only superposition of them. If I want to find all the marked elements, not superposition, I could try this:
1) Do Grover search, get superposition of t marked ele,
2) observe ele space, get one marked ele,
3) remove that ele,
4) goto 1)
This takes time step $O(\sqrt{\frac{N}{t}}+\sqrt{\frac{N-1}{t-1}}+\dots+\sqrt{\frac{N-t+1}{1}})$.
My question is, is this optimal?
