An ensemble of product states is said to exhibit “quantum nonlocality without entanglement” if it cannot be optimally discriminated using local operations and classical communication (LOCC). We show that this property can depend on the chosen discrimination measure. Specifically, we construct a family of ensembles, each consisting of six linearly independent, equally probable bipartite product states, for which LOCC fails to achieve optimal minimum-error discrimination but succeeds in achieving optimal unambiguous discrimination. We further extend our construction to multipartite systems and provide strong numerical evidence that a similar separation between local and global optima is present for minimum-error discrimination, but not for unambiguous discrimination.
We prove limitations on LOCC and separable measurements in bipartite state discrimination problems using techniques from convex optimization. Specific results that we prove include: an exact formula for the optimal probability of correctly discriminating any set of either three or four Bell states via LOCC or separable measurements when the parties are given an ancillary partially entangled pair of qubits; an easily checkable characterization of when an unextendable product set is perfectly discriminated by separable measurements, along with the first known example of an unextendable product set that cannot be perfectly discriminated by separable measurements; and an optimal bound on the success probability for any LOCC or separable measurement for the recently proposed state discrimination problem of Yu, Duan, and Ying.
Any separable measurement with a finite number of outcomes on a finite dimensional multiparticle quantum system can, with non-zero probability, be implemented with local operations and classical communication (LOCC). I use this fact to obtain a necessary and sufficient condition for a finite set of states of a finite dimensional multiparticle quantum system to be amenable to LOCC unambiguous state discrimination. This condition is valid for states which may be be mixed, entangled or both. It is also found to have an intriguing connection with the theory of entanglement witnesses.
We consider the problem of unambiguous (error-free) discrimination of
N linearly independent pure quantum states with prior probabilities, where the goal is to find a measurement that maximizes the average probability of success. We derive an upper bound on the optimal average probability of success using a result on optimal local conversion between two bipartite pure states. We prove that for any $N\ensuremath{\ge}2$ an optimal measurement in general saturates our bound. In the exceptional cases we show that the bound is tight, but not always optimal.
We present a generic study of the unambiguous discrimination between two mixed quantum states. We derive operational optimality conditions and show that the optimal measurements can be classified according to their rank. In Hilbert space dimensions less than or equal to
5, this leads to the complete optimal solution. We demonstrate our method with a physical example, namely, the unambiguous comparison of
n quantum states, and find the optimal success probability.
We consider an infinite class of unambiguous quantum state discrimination problems on multipartite systems, described by Hilbert space
H, of any number of parties. Restricting consideration to measurements that act only on
H, we find the optimal global measurement for each element of this class, achieving the maximum possible success probability of
1/2 in all cases. This measurement turns out to be both separable and unique, and by our recently discovered necessary condition for local quantum operations and classical communication (LOCC) it is easily shown to be impossible by any finite-round LOCC protocol. We also show that, quite generally, if the input state is restricted to lie in
H, then any LOCC measurement on an enlarged Hilbert space is effectively identical to an LOCC measurement on
H. Therefore, our necessary condition for LOCC demonstrates directly that a higher success probability is attainable for each of these problems using general separable measurements as compared to that which is possible with any finite-round LOCC protocol.