How to Solve Problems

Most quantum-mechanics problems become manageable when you separate the modeling choices from the calculation. The common failure mode is to start manipulating formulas before deciding what the system is, what state space is being used, which Hamiltonian generates the dynamics, and which observable is actually being asked about.

This page gives a reusable workflow. It is not a substitute for the canonical pages on states, observables, Hamiltonians, or the Born rule. It is the checklist that keeps those ideas in the right order.

The Core Workflow

1. Identify the physical system.
2. Identify the Hilbert space.
3. Identify the Hamiltonian.
4. Identify the observables.
5. Choose a basis or representation.
6. Solve exactly or approximately.
7. Convert the solution into probabilities, expectation values, or spectra.
8. Check dimensions, limits, normalization, symmetries, and interpretation.

The order matters. You can sometimes shortcut steps in a familiar example, but the shortcuts should be conscious.

Identify the System

Start with the degrees of freedom and the regime of validity. Ask:

• Is the system a particle on a line, a spin, an oscillator, a rigid rotor, a many-particle system, or an effective two-level system?
• Is the treatment nonrelativistic?
• Is particle number fixed?
• Are spin, identical-particle exchange, external fields, or environmental effects relevant?
• Which interactions are included, and which are being idealized away?

For example, a neutral particle in a one-dimensional box, an electron spin in a magnetic field, and an electron in a hydrogenic Coulomb potential are different problems not because the algebra looks different at first, but because the physical systems demand different Hilbert spaces and Hamiltonians.

Identify the Hilbert Space

The Hilbert space tells you what kind of state can exist before you choose coordinates or a basis. Common cases include:

System	Hilbert space
Spinless particle on a line	$L^2(\mathbb R)$, or a subspace fixed by boundary conditions
Particle in a box	$L^2([0,L])$ with the chosen boundary conditions
Spin-$1/2$ system	$\mathbb C^2$
Two qubits	$\mathbb C^2\otimes\mathbb C^2$
Particle with spin	$L^2(\mathbb R^3)\otimes\mathbb C^2$

This step prevents a common mistake: treating a wavefunction, a column vector, and a ket as different theories. They are usually different representations of a state in the same underlying formalism. See Wavefunctions as Representations and Tensor Products for the canonical versions.

Identify the Hamiltonian

The Hamiltonian encodes the energy model and generates time evolution. In a closed system,

$$i\hbar \frac{d}{dt}\lvert \psi(t)\rangle = H\lvert \psi(t)\rangle .$$

For a single nonrelativistic particle in one dimension with potential $V(x)$,

$$H = -\frac{\hbar^2}{2m}\frac{d^2}{dx^2}+V(x).$$

Do not treat this as the universal Hamiltonian. It is a model for a particular regime. A spin in a magnetic field, a two-level atom, a harmonic oscillator, or a charged particle in a vector potential needs a different Hamiltonian.

For wave-mechanics problems, the Hamiltonian is not complete without its domain and boundary conditions. Infinite walls, periodic boundaries, radial regularity, and delta-function interactions can change the allowed states and spectra. The canonical wave-mechanics guide expands this point in How to Solve a Wave-Mechanics Problem.

Identify the Observables

The question usually asks for an observable quantity, not for a state in the abstract. Decide whether you need:

• a probability of a measurement outcome,
• an expectation value,
• a variance or uncertainty,
• an energy spectrum,
• a transition probability,
• a correlation between subsystems,
• a time scale or frequency,
• a scattering reflection or transmission coefficient.

Then translate the request into an operator, projector, POVM element, or probability density. For a normalized pure state and a projective measurement with projector $P_a$, the probability of outcome $a$ is

$$p(a)=\langle \psi\vert P_a\vert\psi\rangle .$$

For a density operator $\rho$, the corresponding trace rule is

$$p(a)=\operatorname{Tr}(\rho P_a).$$

Use Expectation Values and Density Operators when the problem is about averages, mixtures, or subsystems.

Choose a Basis or Representation

Choose the representation that makes the problem simple, not the one that appears first in memory.

If the problem emphasizes	Natural choice
Position-space potentials	position representation
Momentum distributions or free motion	momentum representation
Energy measurements or time-independent dynamics	energy eigenbasis
Spin measurements	eigenbasis of the measured spin component
Tensor-product subsystems	product basis with a declared ordering
Symmetry	simultaneous eigenbasis of commuting symmetry generators

The same state may look simple in one basis and complicated in another. The probabilities are basis-independent only after the state, observable, and inner product are transformed consistently. See Change of Basis and Probability in Different Bases.

Solve the Mathematical Problem

The mathematical task depends on the previous choices:

• Stationary bound-state problems become eigenvalue equations.
• Time-dependent closed-system problems become initial-value problems.
• Scattering problems require asymptotic incoming and outgoing states.
• Spin and qubit problems often reduce to finite-dimensional linear algebra.
• Composite-system problems often require tensor products and partial traces.
• Approximation problems require identifying a small parameter or controlled limit.

For time-independent Hamiltonians, a common strategy is to expand in energy eigenstates:

$$\lvert \psi(0)\rangle =\sum_n c_n\lvert E_n\rangle, \qquad \lvert \psi(t)\rangle =\sum_n c_n e^{-iE_nt/\hbar}\lvert E_n\rangle .$$

For continuous spectra, sums become integrals and normalizable wave packets replace ideal eigenstates. See Normalization and Discrete and Continuous Spectra before manipulating delta-normalized states casually.

Interpret the Result

A mathematical answer is not finished until it has been interpreted in the language of the original physical question. Convert amplitudes into probabilities, check which quantities are directly observable, and state what assumptions the answer depends on.

Useful interpretation questions:

• What measurement would reveal this quantity?
• Does the answer describe a single trial, an ensemble average, or a probability distribution?
• Are the units correct?
• Does the result reduce to the expected answer in a simple limit?
• Does it respect normalization and conservation laws?
• Does it depend on an arbitrary choice of basis, phase, or gauge?
• Is the approximation controlled?

If an answer changes under a harmless global phase or a unitary change of basis, something has probably gone wrong.

Common Problem Types

Bound-State Wave Mechanics

For a one-dimensional bound-state problem, the usual path is:

1. Write $H=-(\hbar^2/2m)d^2/dx^2+V(x)$.
2. Specify the interval and boundary conditions.
3. Solve $H\psi_n=E_n\psi_n$.
4. Normalize $\psi_n$.
5. Expand the initial state in the energy basis if time dependence is needed.
6. Compute probabilities or expectation values from the normalized state.

The Infinite Square Well and Finite Square Well are standard laboratories for this workflow.

Spin and Two-Level Systems

For a two-level system, the usual path is:

1. Choose a basis, such as $\{\lvert 0\rangle,\lvert 1\rangle\}$ or the eigenbasis of $\sigma_z$.
2. Write the state as a normalized two-component vector.
3. Write observables as Hermitian $2\times2$ matrices.
4. Diagonalize the measured observable if needed.
5. Use inner products or projectors to compute probabilities.
6. Use $U(t)=\exp(-iHt/\hbar)$ for closed-system time evolution.

The key discipline is to distinguish the basis used to write the state from the basis associated with the measurement.

Composite Systems

For systems with parts, first declare the tensor-product ordering. For example, in a two-qubit problem decide whether $\lvert 01\rangle$ means first subsystem in $\lvert0\rangle$ and second subsystem in $\lvert1\rangle$. Then ask whether the requested quantity is global or local.

If the question concerns only subsystem $A$, a reduced density matrix is often the cleanest object:

$$\rho_A=\operatorname{Tr}_B(\rho_{AB}).$$

Use the full state for joint measurements and the reduced state for local predictions. Do not infer faster-than-light signaling from entanglement correlations.

Quick Checklist

Before accepting an answer, check:

Check	Question
State	Is the state normalized, or intentionally delta-normalized?
Hamiltonian	Does it match the physical assumptions and units?
Domain	Are boundary conditions or operator domains being ignored?
Observable	Is the measured quantity represented correctly?
Basis	Were states and operators transformed consistently?
Probability	Do probabilities sum or integrate to one?
Expectation value	Is the expectation value real for an observable?
Symmetry	Are conserved quantities and degeneracies respected?
Limit	Does the answer behave correctly as a parameter goes to a simple limit?
Interpretation	Does the final statement answer the original physical question?

Common Mistakes

• Starting with a memorized formula before identifying the system.
• Treating every wavefunction-looking object as normalizable.
• Forgetting that boundary conditions are part of the problem.
• Computing amplitudes but reporting them as probabilities.
• Measuring in one basis while using coefficients from another basis.
• Assuming a state is an eigenstate of the observable being measured.
• Dropping degeneracy in a spectral decomposition or measurement update.
• Confusing a mixed state with a unique unknown pure state.
• Ignoring tensor-product ordering in composite systems.
• Trusting an expression that fails a units, normalization, or limiting-case check.

Where to Go Next

• For readiness checks, use Prerequisites Overview and the Mathematical Toolkit Diagnostic Checklist.
• For the abstract structure of the theory, use the Core Formalism Dependency Graph.
• For coordinate-space problems, use How to Solve a Wave-Mechanics Problem.
• For recurring conceptual traps, use Common Mistakes.
• For compact lookup after you understand the assumptions, use the Reference.

References

• D. J. Griffiths and D. F. Schroeter, Introduction to Quantum Mechanics, 3rd ed., Cambridge University Press, 2018.
• R. Shankar, Principles of Quantum Mechanics, 2nd ed., Springer, 1994.
• J. J. Sakurai and J. Napolitano, Modern Quantum Mechanics, 3rd ed., Cambridge University Press, 2020.
• C. Cohen-Tannoudji, B. Diu, and F. Laloë, Quantum Mechanics, Wiley, 1977.

Exercises

1. A normalized state is written as $\lvert\psi\rangle=(3\lvert0\rangle+4e^{i\phi}\lvert1\rangle)/5$. What extra information is needed before you can compute the probability of a measurement result?

Solution

You need to know which observable is being measured, or equivalently which projectors or measurement basis define the outcomes. In the $\{\lvert0\rangle,\lvert1\rangle\}$ basis, the probabilities are $9/25$ and $16/25$, independent of $\phi$. In another basis, the relative phase can affect the probabilities.

2. A particle is said to be in a box of length $L$, but no boundary condition is stated. Why is the problem incomplete?

Solution

The Hamiltonian differential expression is not enough. The allowed domain fixes which wavefunctions are admissible and which energy eigenvalues occur. Infinite-wall, periodic, Neumann, and mixed boundary conditions lead to different spectra and eigenfunctions.

3. A two-qubit state is $\rho_{AB}$. Which object should you use to predict outcomes of measurements made only on subsystem $A$?

Solution

Use the reduced state $\rho_A=\operatorname{Tr}_B(\rho_{AB})$. For a local observable $O_A$, the expectation value is $\operatorname{Tr}_A(\rho_A O_A)$, equivalently $\operatorname{Tr}_{AB}(\rho_{AB}(O_A\otimes I_B))$.