Abstract
The uncertainty principle is one of the characteristic properties of quantum theory based on incompatibility. Apart from the incompatible relation of quantum states, mutually exclusiveness is another remarkable phenomenon in the information theoretic foundation of quantum theory. We investigate the role of mutual exclusive physical states in the recent work of stronger uncertainty relations for all incompatible observables by Mccone and Pati and generalize the weighted uncertainty relation to the product form as well as their multiobservable analogues. The new bounds capture both incompatibility and mutually exclusiveness, and are tighter compared with the existing bounds.
Introduction
Heisenberg’s uncertainty principle^{1} is one of the fundamental notions in quantum theory. The original form is a result of noncommutativity of the position and momentum operators. Robertson’s formulation of Heisenberg’s uncertainty principle in matrix mechanics^{2} states that for any pair of observables A and B with bounded spectrums, the product of standard deviations of A and B is no less than half of the modulus of the expectation value of their commutator:
where is the standard deviation of the selfadjoint operator A. Here the expectation value is over the state for any observable . In fact, Robertson’s uncertainty relation can be derived from a slightly strengthened Schrödinger uncertainty inequality^{3}
where .
Besides their importance in quantum mechanics, uncertainty relations play a significant role in quantum information theory as well^{4,5,6,7,8,9,10}. The variancebased uncertainty relations possess clear physical meanings and have a variety of applications in quantum information processings such as quantum spin squeezing^{11,12,13,14,15}, quantum metrology^{16,17,18}, and quantum nonlocality^{19,20}.
While the early forms of variancebased uncertainty relations are vital to the foundation of quantum theory, there are two problems still need to be addressed: (1) Homogeneous product of variances may not fully capture the concept of incompatibility. In other words, a weighted relation may produce a better approximation (e.g., the uncertainty relation with Rényi entropy^{21} and variancebased uncertainty relation for a weighted sum), for more details and examples, see ref. 22; (2) The existing variancebased uncertainty relations are far from being tight, and improvement is needed. One also needs to know how to generalize the product form to the case of multiple observables for practical applications.
In ref. 22, the authors and collaborators have proposed weighted uncertainty relations to answer the first question and succeeded in improving the uncertainty relation. Let’s recall the weighted uncertainty relation for the sum of variances. For arbitrary two incompatible observables A, B and any real number λ, the following inequality holds
with
and
where , , and are orthogonal to ψ〉. In informationtheoretic context, it is also natural to quantify the uncertainty by weighted products of variances, which also help to estimate individual variance as in ref. 22.
Recently, Maccone and Pati obtained an amended HeisenbergRobertson inequality^{23}:
which is reduced to HeisenbergRobertson’s uncertainty relation when minimizing the lower bound over , and the equality holds at the maximum. This amended inequality gives rise to a stronger uncertainty relation for almost all incompatible observables, and the improvement is due to the special vector perpendicular to the quantum state ψ〉. We notice that this can be further improved by using the mutually exclusive relation between and ψ〉. Moreover, this idea can be generalized to the case of multiobservables. For this reason the strengthen uncertainty relation thus obtained will be called a mutually exclusive uncertainty relation.
The goal of this paper is to answer the aforementioned questions to derive the product form of the weighted uncertainty relation, and investigate the physical meaning and applications of the mutual exclusive physical states in variancebased uncertainty relations. Moreover, we will generalize the product form to multiobservables to give tighter lower bounds.
Results
We first generalize the weighted uncertainty relations from the sum form^{22} to the product form, and then introduce mutually exclusive uncertainty relations (MEUR). After that we derive a couple of lower bounds based on Mutually exclusive physical states (MEPS), and we show that our results outperform the bound in ref. 23, which has been experimentally tested recently^{24}. Finally, generalization to multiobservables is also given.
We start with the sum form of the uncertainty relation, which takes equal contribution of the variance from each observable. However, almost all variancebased uncertainty relations do not work for the general situation of incompatible observables, and they often exclude important cases. In ref. 22, the authors and collaborators solved this degeneracy problem by considering weighted uncertainty relations to measure the uncertainty in all cases of incompatible observables. Using the same idea, we will study the product form of weighted uncertainty relations to give new and alternative uncertainty relations in the general situation. The corresponding mathematical tool is the famous Young’s inequality. The new weighted uncertainty is expected to reveal the lopsided influence from observables. They contain the usual homogeneous relation of ΔA^{2}ΔB^{2} as a special case.
Theorem 1. Let A, B be two observables such that ΔAΔB > 0, and p, q two real numbers such that . Then the following weighted uncertainty relation for the product of variances holds.
where p < 1, and the equality holds if and only if ΔA = ΔB. If p > 1, then becomes a upper bound for the weighted product.
See Methods for a proof of Theorem 1.
The weighted uncertainty relations for the product of variances have a desirable feature: our measurement of incompatibility is weighted, which fits well with the reality that observables usually don’t always reach equilibrium, i.e., in physical experiments their contributions may not be the same (cf. ref. 22). As an illustration, let us consider the relative error function between the uncertainty and weighted bound, which is defined by
In general f is a function of both p and ψ〉. It is hard to find its extremal points as it involves in partial differential equations. Also the extremal points hardly occur at homogeneous weights, so incompatible observables usually don’t contribute equally to the uncertainty relation, which explains the need for a weighted uncertainty relation in the product form.
In what follows, we show how to tighten Maccone and Pati’s amended HeisenbergRobertson uncertainty relation^{23} by regarding mutually exclusive physical states as another information resource, and then generalize the variancebased uncertainty relation to the case of multiobservables.
We will refer to (6) as a mutually exclusive uncertainty relation since the states ψ〉 and represent two mutual exclusive states in quantum mechanics, which is the main reason for improving the tightness of the bound. Next we move further to improve the bound by combining mutually exclusive relations and weighted relations.
Maccone and Pati’s uncertainty relation can be viewed as a singular case in a family of uncertainty relations parameterized by positive variable λ, which corresponds to our recent work on weighted sum of uncertainty relations^{22}. We proceed similarly as the case of the amended HeisenbergRobertson uncertainty relation by considering a modified squaremodulus and Holevo inequalities in Hilbert space^{25} in the following result.
Theorem 2. Let A and B be two incompatible observables and ψ〉 a fixed quantum state. Then the mutually exclusive uncertainty relation holds:
for any unit vector perpendicular to ψ〉 and arbitrary parameter λ > 0.
See Methods for a proof of Theorem 2.
The obtained variancebased uncertainty relation is stronger than Maccone and Pati’s amended uncertainty relation. In fact, when the maximal value is reached at a point λ_{0} ≠ 1, the new bound is stronger than that of MacconePati’s amended uncertainty relation. Let (i = 1, 2) be two lower bounds given in the RHS of (8), define the tropical sum
This gives a tighter lower bound when the maximal value of is reached at different direction in H_{ψ} (hyperplane orthogonal to ψ〉) for . In other words, the new lower bound is a piecewise defined function of MEPS taking the maximum of the two bounds. In particular, for λ_{0} ≠ 1, the tropical sum offers a better lower bound than , the MacconePati’s lower bound. Note that may have a smaller minimum value than when λ ≠ 1, as , while the minimum value of is just the bound for HeisenbergRobertson’s uncertainty relation. Because we only consider the maximum, it does not affect our result.
For example, consider a 4dimensional system with state , and take the following observables
Direct calculation gives
and
For and , set
and
both of them have modulus one, then
meanwhile
so
Both the lower bounds and are functions of MEPS . However, for each , gives a better approximation of ΔAΔB than . Figure 1 is a schematic diagram of these two lower bounds. It is clear that provides a closer estimate to ΔAΔB:
for any unit MEPS orthogonal to ψ〉. This is due to the fact that the bound is continuous on both MEPS and λ, which shows the advantage of our mutually exclusive uncertainty principle. The shadow region in Fig. 2. illustrates the outline of ΔAΔB and our bound .
In Fig. 3, we illustrate our results, showing how the obtained bound outperforms the recent work of ref. 26 as well as the Schrödinger uncertainty relation. We consider the angular momenta L_{x} and L_{y} for spin1 particle with state ψ〉 = cos θ 1〉 − sin θ 0〉 and = sin θ 1〉 + cos θ 0〉, where 0〉 and 1〉 are eigenstates of the angular momentum L_{z}.
Mutually exclusive physical states with different directions in H_{ψ} offer different kinds of mutually exclusive information and improvement of the uncertainty relation. When such an experiment of the mutually exclusive uncertainty relation is performed, one is expected to have infinitely many strong lower bounds of the variancebased uncertainty relation.
Now we further generalize the uncertainty relations to multiobservables. For simplicity, write
So
is continuous on both MEPS and λ. Repeatedly using (8) for and λ_{jk}, we obtain the following relation.
Theorem 3. Let A_{1}, A_{2}, …, A_{n} be n incompatible observables, ψ〉 a fixed quantum state and λ_{jk} positive real numbers, we have
for any MEPS orthogonal to ψ〉 with modulus one. If some is negative, a negative sign is inserted into the RHS of (14) to ensure positivity. The equality holds if and only if MEPS for all j > k.
As a corollary, Theorem 3 leads to a simple bound of the uncertainty relation for multiobservables.
Corollary 1. Let A_{1}, A_{2}, …, A_{n} be n incompatible observables, then the following uncertainty relation holds
See Methods for a proof of Corollary 1.
Next, we provide yet another mutually exclusive uncertainty relation.
Theorem 4. Let A and B be two incompatible observables and ψ〉 a fixed quantum state. Then
for any unit MEPS orthogonal to ψ〉, with
where MEPS are unit vectors in H_{ψ}.
See Methods for a proof of Theorem 4.
Obviously, (16) can be seen as an amended Schrödinger inequality and also offers a better bound than (2) and MacconePati’s relation (6). Figure 4 illustrates the schematic comparison.
In general, if there exists an operator M for A and B such that 〈M〉 = 0, , then we have the following:
Remark 1. Let A and B be two incompatible observables and ψ〉 a fixed quantum state. We claim the following mutually exclusive uncertainty relation holds:
Eq. (18) also gives a generalized Schrödinger uncertainty relation. Here as usual MEPS is any unit vector perpendicular to ψ〉. The proof of Theorem 4 and Remark 1 are similar to that of Theorem 2, so we sketch it here. It is easy to see that the RHS of (18) reduces to the lower bound of Schrödinger’s uncertainty relation (2) when minimizing over , and the equality holds at the maximum. The corresponding uncertainty relation for arbitrary n observables is the following result.
Theorem 5. Let A_{1}, A_{2}, …, A_{n} be n incompatible observables, ψ〉 a fixed quantum state and λ_{jk} positive real numbers. Then we have that
where M_{jk} satisfy 〈M_{jk}〉 = 0, and MEPS orthogonal to ψ〉 with modulus one.
The RHS of (19) has the minimum value
and the equality holds at the maximum. Therefore one obtains the following corollary.
Corollary 2. Let A_{1}, A_{2}, …, A_{n} be n incompatible observables, then the following uncertainty relation holds
See Methods for a proof of Corollary 2.
We note that our enhanced Schrödinger uncertainty relations offer significantly tighter lower bounds than that of MacconePati’s uncertainty relations for multiobservables, as our lower bound contains an extra term of (compare (1) with (2)).
Finally, we remark that we can also replace the nonhermitian operator in (6) by a hermitian one. A natural consideration is the amended uncertainty relation
for any unit MEPS perpendicular to ψ〉. The corresponding uncertainty relation for multiobservables can also be generalized.
The minimum of Maccone and Pati’s amended bound in the RHS of (6) agrees with the bound in HeisenbergRobertson’s uncertainty relation, which is weaker than Schrödinger’s bound in (2). We point out that the bound given as a continuous function of MEPS’s will always produce a better lower bound. In fact, the continuity of in MEPS implies that there exists suitable such that is tighter than the bound of HeisenbergRobertson’s uncertainty relation. Similarly our lower bound given in (27) or more generally in (18) provides a tighter lower bound than the enhanced Schrödinger’s uncertainty relation (2). This shows the advantage of lower bounds with MEPS’s. Furthermore, lower bounds with more variables give better estimates for the product of variances of observables, as in (19).
Conclusions
The HeisenbergRobertson uncertainty relation is a fundamental principle of quantum theory. It has been recently generalized by Maccone and Pati to an enhanced uncertainty relation for two observables via mutually exclusive physical states. Based on these and weighted uncertainty relations^{22}, we have derived uncertainty relations for the product of variances from mutually exclusive physical states (MEPS) and offered tighter bounds.
In summary, we have proposed generalization of variancebased uncertainty relations. By virtue of MEPS, we have introduced a family of infinitely many Schrödingerlike uncertainty relations with tighter lower bounds for the product of variances. Indeed, our mutually exclusive uncertainty relations can be degenerated to the classical variancebased uncertainty relations by fixing MEPS and the weight. Also, our study further shows that the mutually exclusiveness between states is a promising information resource.
Methods
Proof of Theorem 1. To prove the theorem, we recall Young’s inequality^{27}: for , p < 1 one has that
Note that the righthand side (RHS) may be negative if p < 1. But this can be avoided by using the symmetry of Young’s inequality to get
Thus our bound is nontrivial. We remark that if p > 1, it is directly from the Young’s inequality^{27}
and equality holds in (22) and (23) only when ΔA = ΔB. ■
Proof of Theorem 2. Here we provide two proofs of the proposed mutually exclusive uncertainty relation (8). The first one, based on weighted relations^{22}, is a natural deformation of ref. 23 and is sketched as follows. By maximizing the RHS of (8), we see that the maximum ΔAΔB is achieved when the mutually exclusive physical state (MEPS) . Clearly our uncertainty relation contains (6) as a special case of λ = 1.
The second proof uses geometric property and is preferred because of its mathematical simplicity and also working for the amended HeisenbergRobertson uncertainty relation^{23}. In fact, the RHS of (6), denoted by , is a continuous function of λ and the unit MEPS . By the vector projection, the maximum value ΔAΔB of over the hyperplane of is attained when . Therefore for any λ > 0
where is the RHS of (6). Similarly
for any λ > 0 and the equality holds if λ = 1, which implies (8) and completes the second proof. ■
Proof of Corollary 1. Obviously, taking the minimum of (14) over MEPS implies that
When λ_{jk} = 1 for all j > k, the minimum is . Meanwhile if λ_{jk} and MEPS vary, Eq. (14) provides a family of mutually exclusive uncertainty relations for arbitrary n observables with (24) as the lower bound. ■
Proof of Theorem 4. By the same method used in deriving (8) it follows that , and is
which equals to the lower bound of the Schrödinger uncertainty (2). We can modify g into a function with the same maximum and lower bound as Schrödinger’s uncertainty relation. Note that s ≤ ΔA^{2}ΔB^{2}, then
which is equivalent to (by solving ΔA^{2}ΔB^{2})
for any unit MEPS orthogonal to ψ〉. In fact, let be the RHS of (27). It is easy to see that and
Hence we have the mutually exclusive uncertainty relation appeared in (27). ■
Proof of Corollary 2. Apparently, taking the minimum of (19) over MEPS implies that
with the minimum is . Meanwhile if the MEPS vary, Eq. (19) provides a family of mutually exclusive uncertainty relations for arbitrary n observables with (28) as the lower bound. ■
Additional Information
How to cite this article: Xiao, Y. and Jing, N. Mutually Exclusive Uncertainty Relations. Sci. Rep. 6, 36616; doi: 10.1038/srep36616 (2016).
Publisher’s note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
References
 1.
Heisenberg, W. Über den anschaulichen Inhalt der quantentheoretischen Kinematik und Mechanik. Z. Phys. 43, 172 (1927).
 2.
Robertson, H. P. The Uncertainty Principle. Phys. Rev. 34, 163 (1929).
 3.
Schrödinger, E. Zum Heisenbergschen Unschärfeprinzip. Ber. Kgl. Akad. Wiss. Berlin 24, 296 (1930).
 4.
Busch, P., Heinonen, T. & Lahti, P. J. Heisenberg’s uncertainty principle. Phys. Rep. 452, 155 (2007).
 5.
Busch, P., Lahti, P. & Werner, R. Colloquium: Quantum rootmeansquare error and measurement uncertainty relations. Rev. Mod. Phys. 86, 1261 (2014).
 6.
BiałynickiBirula, I. & Mycielski, J. Uncertainty relations for information entropy in wave mechanics. Comm. Math. Phys. 44, 129 (1975).
 7.
Deutsch, D. Uncertainty in quantum measurements. Phys. Rev. Lett. 50, 631 (1983).
 8.
Wehner, S. & Winter, A. Entropic uncertainty relationsa survey. New J. Phys. 12, 025009 (2010).
 9.
Ghirardi, G. C., Marinatto, L. & Romano, R. An optimal entropic uncertainty relation in a twodimensional Hilbert space. Phys. Lett. A 317, 32 (2003).
 10.
Bosyk, G. M., Portesi, M. & Plastino, A. Collision entropy and optimal uncertainty. Phys. Rev. A 85, 012108 (2012).
 11.
Walls, D. F. & Zoller, P. Reduced quantum fluctuations in resonance. Phys. Rev. Lett. 47, 709 (1981).
 12.
Wódkiewicz, K. & Eberly, J. Coherent states, squeezed fluctuations, and the SU(2) am SU(1,1) groups in quantumoptics applications. J. Opt. Soc. Am. B 2, 458 (1985).
 13.
Wineland, D. J., Bollinger, J. J., Itano, W. M., Moore, F. L. & Heinzen, D. J. Spin squeezing and reduced quantum noisd in spectroscopy. Phys. Rev. A 46 R6797 (1992).
 14.
Kitagawa, M. & Ueda, M. Squeezed spin states. Phys. Rev. A 47, 5138 (1993).
 15.
Ma, J., Wang, X. G., Sun, C. P. & Nori, F. Quantum spin squeezing. Phys. Rep. 509, 89 (2011).
 16.
Giovannetti, V., Lloyd, S. & Maccone, L. Quantumenhanced measurements: beating the standard quantum limit. Science 306, 1330 (2004).
 17.
Giovannetti, V., Lloyd, S. & Maccone, L. Quantum metrology. Phys. Rev. Lett. 96, 010401 (2006).
 18.
Giovannetti, V., Lloyd, S. & Maccone, L. Advances in quantum metrology. Nat. Photon. 5, 222 (2011).
 19.
Oppenheim, J. & Wehner, S. The uncertainty principle determines the nonlocality of quantum mechanics. Science. 330, 1072 (2010).
 20.
Li, J.L., Du, K. & Qiao, C.F. Ascertaining the uncertainty relations via quantum correlations. J. Phys. A 47, 085302 (2014).
 21.
Maassen, H. & Uffink, J. B. M. Generalized entropic uncertainty relations. Phys. Rev. Lett. 60, 1103 (1988).
 22.
Xiao, Y., Jing, N., LiJost, X. & Fei, S.M. Weighted uncertainty relations. Sci. Rep. 6, 23201 (2016).
 23.
Maccone, L. & Pati, A. K. Stronger uncertainty relation for all incompatible observables. Phys. Rev. Lett. 113, 260401 (2014).
 24.
Wang, K., Zhan, X., Bian, Z., Li, J., Zhang, Y. & Xue, P. Experimental investigation of the stronger uncertainty relations for all incompatible observables. Phys. Rev. A. 93, 052108 (2016).
 25.
Holevo, A. S. A generalization of the RaoCramér inequality. Teor. Veroyatnost. i Primenen. 18, 371 (1973) Probab. Appl. 18, 359 (1974).
 26.
Mondal, D., Bagchi, S. & Pati, A. K. Tighter uncertainty and reverse uncertainty relations. arXiv: 1607.06712v1.
 27.
Zorich, V. A. Mathematical Analysis 1. (Springer, Berlin, 2015).
Acknowledgements
We thank Jian Wang, Yinshan Chang, Xianqing LiJost and ShaoMing Fei for fruitful discussions. The work is supported by National Natural Science Foundation of China (Grants Nos 11271138 and 11531004), China Scholarship Council and Simons Foundation Grant No. 198129.
Author information
Affiliations
School of Mathematics, South China University of Technology, Guangzhou 510640, China
 Yunlong Xiao
 & Naihuan Jing
Max Planck Institute for Mathematics in the Sciences, Leipzig 04103, Germany
 Yunlong Xiao
Department of Mathematics, North Carolina State University, Raleigh, NC 27695, USA
 Naihuan Jing
Authors
Search for Yunlong Xiao in:
Search for Naihuan Jing in:
Contributions
Y.X. and N.J. analyzed and wrote the manuscript.
Competing interests
The authors declare no competing financial interests.
Corresponding author
Correspondence to Naihuan Jing.
Rights and permissions
This work is licensed under a Creative Commons Attribution 4.0 International License. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in the credit line; if the material is not included under the Creative Commons license, users will need to obtain permission from the license holder to reproduce the material. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/
About this article
Further reading

1.
On noiseresolution uncertainty in quantum field theory
Scientific Reports (2017)

2.
A Stronger Multiobservable Uncertainty Relation
Scientific Reports (2017)
Comments
By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.