Abstract
In the study of local discrimination for multipartite unitary operations, Duan et al. (Phys Rev Lett 100(2):020503, 2008) exhibited an ingenious expression: Any two different unitary operations \(U_1\) and \(U_2\) are perfectly distinguishable by local operations and classical communication in the single-run scenario if and only if 0 is in the local numerical range of \(U_1^\dag U_2\). However, how to determine when 0 is in the local numerical range remains unclear. So it is generally hard to decide the local discrimination of nonlocal unitary operations with a single run. In this paper, for two-qubit diagonal unitary matrices V and their local unitary equivalent matrices, we present a necessary and sufficient condition for determining whether the local numerical range is a convex set or not. The result can be used to easily judge the locally perfect distinguishability of any two unitary operations \(U_1\) and \(U_2\) satisfying \(U_1^\dag U_2=V\). Moreover, we design the corresponding protocol of local discrimination. Meanwhile, an interesting phenomenon is discovered: Under certain conditions with a single run, \(U_1\) and \(U_2\) such that \(U_1^\dag U_2=V\) are locally distinguishable with certainty if and only if they are perfectly distinguishable by global operations.
Similar content being viewed by others
References
Sacchi, M.F.: Optimal discrimination of quantum operations. Phys. Rev. A 71(6), 062340 (2005)
Wang, G.M., Ying, M.S.: Unambiguous discrimination among quantum operations. Phys. Rev. A 73(4), 042301 (2006)
Duan, R.Y., Feng, Y., Ying, M.S.: Perfect distinguishability of quantum operations. Phys. Rev. Lett. 103(21), 210501 (2009)
Ji, Z.F., Feng, Y., Duan, R.Y., Ying, M.S.: Identification and distance measures of measurement apparatus. Phys. Rev. Lett. 96(20), 200401 (2006)
Sedlak, M., Ziman, M.: Optimal single-shot strategies for discrimination of quantum measurements. Phys. Rev. A 90(5), 052312 (2014)
Mikova, M., Sedlak, M., Straka, I., Micuda, M., Ziman, M., Jezek, M., Dusek, M., Fiurasek, J.: Optimal entanglement-assisted discrimination of quantum measurements. Phys. Rev. A 90(2), 022317 (2014)
Cao, T.Q., Gao, F., Zhang, Z.C., Yang, Y.H., Wen, Q.Y.: Perfect discrimination of projective measurements with the rank of all projectors being one. Quantum Inf. Process. 14, 2645–2656 (2015)
Piani, M., Watrous, J.: All entangled states are useful for channel discrimination. Phys. Rev. Lett. 102(25), 250501 (2009)
Matthews, W., Piani, M., Watrous, J.: Entanglement in channel discrimination with restricted measurements. Phys. Rev. A 82(3), 032302 (2010)
Childs, A.M., Preskill, J., Renes, J.: Quantum information and precision measurement. J. Mod. Opt. 47(2–3), 155–176 (2000)
Acín, A.: Statistical distinguishability between unitary operations. Phys. Rev. Lett. 87(17), 177901 (2001)
D’Ariano, G.M., Presti, P.L., Paris, M.G.A.: Using entanglement improves the precision of quantum measurements. Phys. Rev. Lett. 87(27), 270404 (2001)
Chefles, A., Sasaki, M.: Retrodiction of generalized measurement outcomes. Phys. Rev. A 67(3), 032112 (2003)
Duan, R.Y., Feng, Y., Ying, M.S.: Entanglement is not necessary for perfect discrimination between unitary operations. Phys. Rev. Lett. 98(10), 100503 (2007)
Chefles, A., Kitagawa, A., Takeoka, M., Sasaki, M., Twamley, J.: Unambiguous discrimination among oracle operators. J. Phys. A Math. Theor. 40, 10183–10217 (2007)
Wu, X.D., Duan, R.Y.: Exact quantum search by parallel unitary discrimination schemes. Phys. Rev. A 78(1), 012303 (2008)
Chiribella, G., DAriano, G.M., Perinotti, P.: Quantum circuit architecture. Phys. Rev. Lett. 101(6), 060401 (2008)
Zhou, X.F., Zhang, Y.S., Guo, G.C.: Unitary transformations can be distinguished locally. Phys. Rev. Lett. 99(17), 170401 (2007)
Duan, R.Y., Feng, Y., Ying, M.S.: Local distinguishability of multipartite unitary operations. Phys. Rev. Lett. 100(2), 020503 (2008)
Li, L.Z., Qiu, D.W.: Local entanglement is not necessary for perfect discrimination between unitary operations acting on two qudits by local operations and classical communication. Phys. Rev. A 77(3), 032337 (2008)
Gawron, P., Puchała, Z., Miszczak, J.A., Skowronek, Ł., Życzkowski, K.: Restricted numerical range: a versatile tool in the theory of quantum information. J. Math. Phys. 51(10), 102204 (2010)
Puchała, Z., Gawron, P., Miszczak, J.A., Skowronek, Ł., Choi, M.D., Życzkowski, K.: Product numerical range in a space with tensor product structure. Linear Algebra Appl. 434, 327–342 (2011)
Acknowledgments
This work is supported by the National Natural Science Foundation of China (Grant Nos. 61272057, 61170270 and 61572081), the Beijing Higher Education Young Elite Teacher Project (Grant Nos. YETP0475 and YETP0477) and the Natural Science Foundation of Shaanxi Province of China (No. 2015JM6263).
Author information
Authors and Affiliations
Corresponding author
Appendix
Appendix
Here Lemma 6 of the main text will be proved.
Firstly, we show the sufficiency. It suffices to show that if \(\phi <\theta <\psi \), then for any \(y=a_1+a_2\hbox {e}^{i\phi }+a_3\hbox {e}^{i\psi }+a_4\hbox {e}^{i\theta }\in W(V)\) with \(a_i\ge 0\) and \(\sum \nolimits _{i=1}^4a_i=1\), there always exist \(s,t\in [0,1]\) such that
As to \(\psi <\theta <\phi \), the proof is similar.
By the equal of the complex numbers, we can obtain
They can be regarded as an equation about s and t.
Further, we have
where
Case 1 If \(A=0\), that is \(\psi =\phi +\theta \), then \(Bt=C\), which induces \(\theta =\pi \) or \(t=a_1+a_3\).
Let \(\theta =\pi \), then \(A=B=C=0\), which implies \(s-t=a_2-a_3\), \(s+t=1+a_1-a_4\). Thus \(s=a_1+a_2,t=a_1+a_3\).
Let \(t=a_1+a_3\), then \(B\ne 0\) which induces
Since \(0<\phi <\theta <\psi <2\pi \) and \(0<\psi =\phi +\theta <2\pi \), then \(\sin \frac{\phi -\theta }{2}<0,\sin \frac{\phi +\theta }{2}>0\). Further
Therefore \(0\le s\le 1\). Now, we have shown that if \(A=0\), then \(s,t\in [0,1]\).
Case 2 Suppose that \(A\ne 0\), then \(s={(C-Bt)}/{A}\). Further we can get that
where
If \(0<\phi <\theta <\psi <2\pi \), then \(c\ge 0\) and \(a+b+c\le 0\). But both cannot be 0 at the same time, otherwise \(\sum \nolimits _{i=1}^4a_i=0\), which is a contradiction.
Case 2.1 Let \(a=0\). Since \(c\ge 0\), \(a+b+c\le 0\), and they cannot be 0 simultaneously; thus, \(t=-c/b\in [0,1]\).
Case 2.2 Let \(a\ne 0\), it need to be divided into three cases:
(I) \(a>0,b\le 0,c\ge 0\); (II) \(a<0,b\ge 0,c\ge 0\); (III) \(a<0,b\le 0,c\ge 0\).
As to the cases (II) (III), \(b^2-4ac\ge 0\). And for (I), \(b^2-4ac=(d-a-c)^2-4ac\ge {(a-c)^2}\ge 0\).
In the following we denote \(t_1=(-b+\sqrt{b^2-4ac})/2a\) and \(t_2=(-b-\sqrt{b^2-4ac})/2a\). Further we will show that \(t=t_2\) is the only solution in [0, 1].
According to (I)–(III), \(t_2\ge 0\). Since \(a+b+c\le 0\), we know that
if \(a>0\), \(a^2+ab+ac\le 0\), then \((2a+b)^2\le b^2-4ac\). If \(2a+b\le \sqrt{b^2-4ac}\), then \(t_1\ge 1\). If \(-2a-b\le \sqrt{b^2-4ac}\), then \(t_2\le 1\).
if \(a<0\), then \(a^2+ab+ac\ge 0\), and thus \((2a+b)^2\ge b^2-4ac\). Again, \(2a+b=d+a-c<0\), then \(t_2\le 1\).
In a word, if \(A\ne 0\), there always exists a solution \(t=-c/b\) or \(t=t_2\) in [0, 1].
Similarly, when \(0<\phi <\theta <\psi <2\pi \), there is also a solution s in [0, 1]. Note that if \(B\ne 0\), we can get a equation about s,
where
Now we turn to consider the necessity. We only need to show that when \(\phi <\psi <\theta , \psi <\phi <\theta , \theta <\phi <\psi \) and \(\theta <\psi <\phi \), \(W^{\mathrm{local}}(V)\subset W(V)\). Namely, there always exists a point, such as \(\frac{1}{2}+\frac{1}{2}\hbox {e}^{i\theta }\), such that \(\frac{1}{2}+\frac{1}{2}\hbox {e}^{i\theta }\in W(V)\), but \(\frac{1}{2}+\frac{1}{2}\hbox {e}^{i\theta }\notin W^{\mathrm{local}}(V)\). First we will prove that when \(0<\phi <\psi <\theta <2\pi \), \(\frac{1}{2}+\frac{1}{2}\hbox {e}^{i\theta }\notin W^{\mathrm{local}}(V)\). And if \(0<\theta <\psi <\phi <2\pi \), the proof is similar.
When \(0<\phi <\psi <\theta <2\pi \), \(0<\frac{\phi -\psi +\theta }{2}<\pi \). So \(A>0\). Thus the equation about t must be obtained.
where
And \(c>0,a+b+c>0\).
Case 1 If \(a=0\), then \(bt=-c\). When \(0<\phi <\psi <\theta <2\pi \), \(a=0\) is equivalent to \(\psi +\theta =\phi +2\pi \). So \(b=4\sin \frac{\psi }{2}\cos \frac{\psi }{2}\sin ^2\frac{\theta }{2}\). Thus b can be positive, negative and zero. Again \(b+c>0,c>0\), it is easy to see that if \(b>0,t=-c/b<0\); if \(-c<b<0, t=-c/b>1\); if \(b=0\), there is no t.
Therefore, if \(a=0\), there is no solution t in [0, 1].
Case 2 Let \(a\ne 0\), it can be divided into four cases:
(I) \(a>0,b\ge 0,c>0\); (II) \(a>0,b\le 0,c>0\);
(III) \(a<0,b\ge 0,c>0\); (IV) \(a<0,b\le 0,c>0\).
The notations \(t_1=(-b+\sqrt{b^2-4ac})/2a\) and \(t_2=(-b-\sqrt{b^2-4ac})/2a\) will also be used as follows.
Case 2.1 As to (III) (IV), \(b^2-4ac>0\) and only \(t_2>0\). Since \(a+b+c>0\) and \(a<0\), then \((2a+b)^2<b^2-4ac\). So \(-2a-b<\sqrt{b^2-4ac}\), thus \(t_2>1\). Therefore, for (III) (IV), there is no solution t in [0, 1].
As to (I) (II), if \(b^2-4ac<0\), there is no t. So let \(b^2-4ac\ge 0\). For (I), there is no solution t in [0, 1]. For (II), since \(a+b+c>0\) and \(a>0\), then \((2a+b)^2>b^2-4ac\). If \(2a+b<0\), then \(t_2>1\), which does not meet our requirement. If \(2a+b>0\), then \(t_1<1\).
For (I) (III) (IV), we have shown that there is no solution t in [0, 1]. For (II), we will suppose that both s and t have the solution in [0, 1] and find the contradiction as follows.
Case 2.2 If \(a>0,b\le 0,c>0\). Assume that \(b^2-4ac\ge 0\) and \(2a+b>0\), then \(t=t_1\in [0,1]\) and \(a>c\). And let \(0\le s\le 1\), where \(s={(C-Bt)}/{A}, A>0, B=a>0\), and \(C=c+d=a+b+2c>0\). So \(C-A\le Bt\le C\). By \(Bt\le C\), then \(\sqrt{{{b}^{2}}-4ac}\le 2a+3b+4c\). Since \(2a+3b+4c\ge 0\), then \(0\ge b\ge (-2a-4c)/3\), thus \({{b}^{2}}-4ac\le 4(a-4c)(a-c)/9<0\), which is due to that \(a-c>0\) and
This is a contradiction.
Therefore, when \(a\ne 0\), not both s and t have the solution in [0, 1]. That is at least one of s and t has no solution in [0, 1].
Above all, if \(0<\phi <\psi <\theta <2\pi \), then \(\frac{1}{2}+\frac{1}{2}\hbox {e}^{i\theta }\in W(V)\), but \(\frac{1}{2}+\frac{1}{2}\hbox {e}^{i\theta }\notin W^{\mathrm{local}}(V)\). That is \(W^{\mathrm{local}}(V)\subset W(V)\).
Similarly, when \(0<\psi <\phi <\theta <2\pi \) and \(0<\theta <\phi <\psi <2\pi \), \(\frac{1}{2}+\frac{1}{2}\hbox {e}^{i\theta }\in W(V)\) and \(\frac{1}{2}+\frac{1}{2}\hbox {e}^{i\theta }\notin W^{\mathrm{local}}(V)\). But it should be noted that in these two situations we must consider the equation about s,
where
To sum up, we have shown that \(W^{\mathrm{local}}(V)=W(V)\) if and only if \(\phi <\theta <\psi \) or \(\psi <\theta <\phi \). \(\square \)
Rights and permissions
About this article
Cite this article
Cao, TQ., Gao, F., Yang, YH. et al. Determination of locally perfect discrimination for two-qubit unitary operations. Quantum Inf Process 15, 529–549 (2016). https://doi.org/10.1007/s11128-015-1175-x
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11128-015-1175-x