In the multitarget Grover algorithm, we are given an unstructured N-element list of objects S_i containing a T-element subset tau and function f, called an oracle, such that f(S_i)=1 if S_i is in tau, otherwise f(S_i) = 0. By using quantum parallelism, an element of tau can be retrieved in O(sqrt(N/T)) steps, compared to O(N/T) for any classical algorithm. It is shown here that in combinatorial optimization problems with N=4^M, the density of states can be used in conjunction with the

more »

... t Grover algorithm to construct a sequence of oracles that structure the search so that the optimum state is obtained with certainty in O(log(N)) steps.