In this chapter, the simplest quantum measurement situation, i.e., where the state is an eigenvector of the observable to be measured, is considered. The states are therefore non-superimposed and measurement Axiom 3.2 of Section 13.7 in the previous chapter is applicable. We consider states in https://www.w3.org/1998/Math/MathML" display="inline"> ℂ 2 , ℂ 2 ⊗ ℂ 2 https://www.w3.org/1999/xlink" xlink:href=" https://s3-euw1-ap-pe-df-pch-content-public-p.s3.eu-west-1.amazonaws.com/9781003264569/7139ce19-0d58-4a0c-820f-b76fb033e438/content/math14_1.tif "/> and https://www.w3.org/1998/Math/MathML" display="inline"> ℂ 2 ⊗ ℂ 2 ⊗ ℂ 2 https://www.w3.org/1999/xlink" xlink:href=" https://s3-euw1-ap-pe-df-pch-content-public-p.s3.eu-west-1.amazonaws.com/9781003264569/7139ce19-0d58-4a0c-820f-b76fb033e438/content/math14_2.tif "/> . The generalisation to https://www.w3.org/1998/Math/MathML" display="inline"> ⊗ n ℂ 2 https://www.w3.org/1999/xlink" xlink:href=" https://s3-euw1-ap-pe-df-pch-content-public-p.s3.eu-west-1.amazonaws.com/9781003264569/7139ce19-0d58-4a0c-820f-b76fb033e438/content/math14_3.tif "/> is then clear.
The representation of physical observables by Hermitian linear operators was postulated assuming that an expectation value, defined by Eqs.(2.11) and (2.12) , is associated with any observable A measured in a quantum state Ψ of a dynamical system. In view of measurement uncertainties, defined by the dispersion: ( Δ A ) 2 = 〈 A 2 〉 − 〈 A 〉 2 = 〈 ( A − 〈 A 〉 ) 2 〉 = 〈 ( δ A ) 2 〉 https://s3-euw1-ap-pe-df-pch-content-public-p.s3.eu-west-1.amazonaws.com/9781003078265/e77b8ce0-ccb3-48ed-9f6d-a84c8cb529c5/content/eqn3_1_B.tif"/> the expectation value 〈 A 〉 is taken in a state Ψ if and only if (Δ Α ) 2 vanishes. It is convenient to introduce the deviation operator: 3.1 δ  =  − 〈 A 〉 Î https://s3-euw1-ap-pe-df-pch-content-public-p.s3.eu-west-1.amazonaws.com/9781003078265/e77b8ce0-ccb3-48ed-9f6d-a84c8cb529c5/content/eqn3_2_B.tif"/> such that the condition of a vanishing dispersion for the observable A in a given state Ψ reads: 〈 Ψ |( δ  ) 2 Ψ 〉 = 0 https://s3-euw1-ap-pe-df-pch-content-public-p.s3.eu-west-1.amazonaws.com/9781003078265/e77b8ce0-ccb3-48ed-9f6d-a84c8cb529c5/content/eqn3_3_B.tif"/> Because δ is a Hermitian operator, we can also write: 〈 δ Â Ψ | δ Â Ψ 〉 = | δ Â Ψ | 2 = 0 https://s3-euw1-ap-pe-df-pch-content-public-p.s3.eu-west-1.amazonaws.com/9781003078265/e77b8ce0-ccb3-48ed-9f6d-a84c8cb529c5/content/eqn3_4_B.tif"/> which means that the norm of δ Â Ψ is zero and this implies that we have a null function δ Â Ψ = 0. In other words: 3.2 Â Ψ = 〈 A 〉 Ψ https://s3-euw1-ap-pe-df-pch-content-public-p.s3.eu-west-1.amazonaws.com/9781003078265/e77b8ce0-ccb3-48ed-9f6d-a84c8cb529c5/content/eqn3_5_B.tif"/> It follows that the state function Ψ where a definite expectation value 〈 A 〉 of the observable A is observed, should be a solution of Eq.(3.2) rather than an arbitrary 112 function. The action of the operator  on Ψ has to be simply that of rescaling. All the state functions which are solutions of Eq.(3.2) are called eigenfunctions of  , and the constant factor 〈 A 〉 is said to be the corresponding eigenvalue. Hence we have shown that if ((δ  ) 2 ) = 0, the state function should be an eigenfunction of  , which means that a physical observable only takes definite values in states represented by eigenfunctions of the associated quantum operator. Conversely, given a quantum operator  one can systematically determine all its eigenvalues and eigenfunctions, which are solutions of the so-called eigenvalue equation: 3.3 Â Ψ = λ Ψ https://s3-euw1-ap-pe-df-pch-content-public-p.s3.eu-west-1.amazonaws.com/9781003078265/e77b8ce0-ccb3-48ed-9f6d-a84c8cb529c5/content/eqn3_6_B.tif"/> It is visible that the eigenvalue equation fixes the eigenfunctions only up to an overall scale factor, which may always be chosen such that the eigenfunctions be considered normalized. Then, in view of Eq.(2.11) , Eq.(3.3) gives: 〈  〉 = ∫ − ∞ ∞ Ψ * Â Ψ d x = λ ∫ − ∞ ∞ Ψ * Ψ d x = λ https://s3-euw1-ap-pe-df-pch-content-public-p.s3.eu-west-1.amazonaws.com/9781003078265/e77b8ce0-ccb3-48ed-9f6d-a84c8cb529c5/content/eqn3_7_B.tif"/> If we further apply the operator  to Eq.(3.3) , we obtain  2 Ψ = λ 2 Ψ and this leads to: 〈  2 〉 = ∫ − ∞ ∞ Ψ *  2 Ψ d x = λ 2 ∫ − ∞ ∞ Ψ * Ψ d x = λ 2 https://s3-euw1-ap-pe-df-pch-content-public-p.s3.eu-west-1.amazonaws.com/9781003078265/e77b8ce0-ccb3-48ed-9f6d-a84c8cb529c5/content/eqn3_8_B.tif"/> It follows that: ( Δ A ) 2 = 〈  2 〉 − 〈  〉 2 = 0 https://s3-euw1-ap-pe-df-pch-content-public-p.s3.eu-west-1.amazonaws.com/9781003078265/e77b8ce0-ccb3-48ed-9f6d-a84c8cb529c5/content/eqn3_9_B.tif"/> and this shows that the dispersion vanishes if a physical observable is measured in states represented by eigenfunctions of the associated operator.
The disturbance during a measurement on a classical system can be made arbitrarily small. One may then assume that a measurement does not change the state, e.g., the radar used by police to measure the speed of a car has a negligible effect on the speed of the car. Classical dynamics also applies to the measuring process as well. For quantum systems Postulate 26.1(OV) tells us that a measurement of a discrete observable A will yield an eigenvalue of its corresponding operator  . Let the initial state vector be ϕ → i https://s3-euw1-ap-pe-df-pch-content-public-p.s3.eu-west-1.amazonaws.com/9780203702413/2f242b9b-104e-4a03-b478-d3f6577cf44e/content/eq6175.tif"/> at time t = 0 when the measurement begins and let the final state vector be ϕ → f https://s3-euw1-ap-pe-df-pch-content-public-p.s3.eu-west-1.amazonaws.com/9780203702413/2f242b9b-104e-4a03-b478-d3f6577cf44e/content/eq6176.tif"/> at time t = Δ t when the measurement ends. A natural question to ask is Can disturbance during a measurement on the system be made arbitrarily small so that the state can remain the same, and if not, does the initial state evolve into the final state in a unitary manner describable in terms of Postulate 29.1.2(TESP) ?
The nature of measurements, and their place within quantum theory, has engaged physicists and philosophers for generations [Bra92; Sch03]. Much of that interest centered on variables such as position and momentum of free particles, whose values form a continuum. The present monograph deals with discrete quantum states; the measurements are those required to specify as completely as possible a particular discrete quantum state Ψ or, more generally, a density matrix ρ defined within a finite-dimensional Hilbert space. General remarks General system . At the outset we assume that the possible quantum states are a small number – the N essential states used in formulating the time-dependent Schrödinger equation or specifying the dimensions of the density matrix. To completely characterize a density matrix for such a system we require the N 2 elements. Of these the N diagonal elements are real valued, while the off-diagonal elements of the upper right side are complex conjugates of those on the lower left. Thus with allowance for the requirement of unit trace a total of N 2 – 1 real numbers suffice to completely specify the density matrix. These values must be consistent with the constraints discussed in Sec. 16.6.3. Pure state . If it is known that the system is in a pure state, we require the magnitude and phase of N probability amplitudes. These are constrained by normalization, and so only 2 N –1 real numbers are needed. Out of these 2 N – 1 parameters the overall phase of the statevector is usually not of interest (but see Chap. 20).
In Chapter 15 , quantum emulations of some familiar invertible digital gates were discussed; specifically the not , cnot , Toffoli and Peres gates, each of which is significant in quantum computation, were considered. The emulations were defined on basis elements only – i.e., on https://www.w3.org/1998/Math/MathML" display="inline"> B ℂ 2 https://www.w3.org/1999/xlink" xlink:href=" https://s3-euw1-ap-pe-df-pch-content-public-p.s3.eu-west-1.amazonaws.com/9781003264569/7139ce19-0d58-4a0c-820f-b76fb033e438/content/math16_3.tif "/> or https://www.w3.org/1998/Math/MathML" display="inline"> B ℂ 2 ⊗ ℂ 2 https://www.w3.org/1999/xlink" xlink:href=" https://s3-euw1-ap-pe-df-pch-content-public-p.s3.eu-west-1.amazonaws.com/9781003264569/7139ce19-0d58-4a0c-820f-b76fb033e438/content/math16_4.tif "/> or https://www.w3.org/1998/Math/MathML" display="inline"> B ℂ 2 ⊗ ℂ 2 ⊗ ℂ 2 https://www.w3.org/1999/xlink" xlink:href=" https://s3-euw1-ap-pe-df-pch-content-public-p.s3.eu-west-1.amazonaws.com/9781003264569/7139ce19-0d58-4a0c-820f-b76fb033e438/content/math16_5.tif "/> etc. On these restricted inputs the quantum emulations were seen to produce outputs identical to their digital equivalents. Unlike the digital gates the quantum emulations may be extended beyond the basis sets to the entire tensor product spaces so defined – i.e., to https://www.w3.org/1998/Math/MathML" display="inline"> ℂ 2 , ℂ 2 ⊗ ℂ 2 , ℂ 2 ⊗ ℂ 2 ⊗ ℂ 2 https://www.w3.org/1999/xlink" xlink:href=" https://s3-euw1-ap-pe-df-pch-content-public-p.s3.eu-west-1.amazonaws.com/9781003264569/7139ce19-0d58-4a0c-820f-b76fb033e438/content/math16_6.tif "/> etc. Such generalised inputs to the quantum gates lead to more general quantum outputs to which the simple deterministic quantum measurement processes of Chapter 15 are not applicable. The quantum gates defined on the extended domains are shown to be unitary, and are denoted in the same way as their restrictions to the basis vectors. We first consider a sufficient condition for an operator to be unitary; it has applications in our discussions below and in Chapter 17 .
A transition would lead the system from its original superposition state to another state, in which the system and the measuring apparatus acquire definite properties. The problem of measurement is approached from the theory of environment-induced decoherence. This theoretical program, based on the study of the effects of the interaction between a quantum system, is considered as an open system and its environment. In turn, even the measurement problem has been reformulated, expanding its scope. As Schlosshauer points out, the conceptual difficulties of quantum measurement can now be concentrated around two cases: The definite outcome, which consists of the traditional question about quantum measurement. The problem of the preferred basis: since there is a theoretical ambiguity in the definition of the observable measured due to the mathematical possibility of a base change in the expression of the system state.
Vector spaces are the mathematical structures in which the state vectors of quantum computation live. When endowed with an inner product, a vector space becomes an inner product space. Of particular relevance are complex inner product spaces. An objective of this chapter is to provide a formal definition of a vector space and give several examples. You will see that, in this context, a ‘vector’ is a much more general object than vectors such as those in https://www.w3.org/1998/Math/MathML" display="inline"> ℝ 2 https://www.w3.org/1999/xlink" xlink:href=" https://s3-euw1-ap-pe-df-pch-content-public-p.s3.eu-west-1.amazonaws.com/9781003264569/7139ce19-0d58-4a0c-820f-b76fb033e438/content/math6_1.tif "/> and https://www.w3.org/1998/Math/MathML" display="inline"> ℝ 3 https://www.w3.org/1999/xlink" xlink:href=" https://s3-euw1-ap-pe-df-pch-content-public-p.s3.eu-west-1.amazonaws.com/9781003264569/7139ce19-0d58-4a0c-820f-b76fb033e438/content/math6_2.tif "/> . However, the already familiar operations of vector addition and scalar multiplication are essential ingredients of all vector spaces. We demonstrate how vectors in a vector space can be constructed using a set of vectors called a basis and see how vectors can be written in terms of different bases. Of particular importance are orthonormal bases in which the basis vectors are mutually orthogonal and normalised. Finally, we introduce cosets and explain how a quotient space is constructed.