1 Hamiltonian in Grover’s Algorithm

1.1 Construction of the Hamiltonian

The Hamiltonian Ĥ is constructed to encode the search problem as an eigenvalue problem in
quantum mechanics. It is designed such that the evolution of the quantum state under Ĥ amplifies
the amplitude of the marked (or target) state. It typically has two components:

Ĥ = Ĥoracle + Ĥdiffusion

1.1.1 Oracle Hamiltonian Ĥoracle

The Oracle Hamiltonian encodes the marked state |w⟩ by introducing a phase or energy penalty:

Ĥoracle = |w⟩⟨w|
This ensures that the marked state has a unique eigenvalue, distinguishing it from other states.

1.1.2 Diffusion Hamiltonian Ĥdiffusion

The Diffusion Hamiltonian drives the system towards the marked state by coupling all states in the
superposition:

Ĥdiffusion = I − |s⟩⟨s|
where |s⟩ is the equal superposition state of all database entries:
N −1
1 X
|s⟩ = √ |i⟩
N i=0
Together, these components guide the quantum state from the initial superposition |s⟩ to the
target |w⟩ by evolving the system under Schrödinger’s equation.

1.2 Role of the Marked State in Defining the Hamiltonian’s Struc-

ture
The marked state |w⟩ defines the eigenstate of the oracle Hamiltonian with a distinct eigenvalue,
often chosen to be different from all other states. This difference introduces a phase shift (or energy
bias) to the marked state, which is crucial for distinguishing it during the quantum evolution. In
the Ĥoracle , |w⟩ has a higher ”weight” or ”priority,” ensuring the system evolves towards amplifying
|w⟩.

1.3 Connection to Schrödinger’s Equation

1.3.1 Time Evolution in Grover’s Algorithm
The time evolution of the quantum state under Schrödinger’s equation in the continuous-time
formalism is governed by:

∂
iℏ |ψ(t)⟩ = Ĥ|ψ(t)⟩
∂t
The Hamiltonian Ĥ determines the rate and direction of this evolution. In Grover’s algorithm:

1
 • The system starts in the equal superposition |s⟩.

• Under Ĥ, the quantum state evolves towards |w⟩ through oscillatory dynamics.

The discrete steps in the standard Grover’s algorithm (oracle and diffusion operations) corre-
spond to small time intervals (∆t) in the continuous-time evolution:

|ψ(t + ∆t)⟩ ≈ e−iĤ∆t/ℏ |ψ(t)⟩

√ total evolution time T required to maximize the probability of finding |w⟩ is proportional
The
to O( N ), matching the step complexity of the discrete Grover algorithm.

1.4 Assumptions in the Derivation of the Algorithm

• Initial State Assumption: The system starts in an equal superposition of all states:
N −1
1 X
|ψ(0)⟩ = |s⟩ = √ |i⟩
N i=0

• Energy Scale Assumption: The Hamiltonian’s eigenvalues are scaled such that the marked
state |w⟩ is distinguishable with minimal energy gap.

• Two-Level Approximation: Only the subspace spanned by |s⟩ (initial state) and |w⟩
(marked state) is considered, reducing the dynamics to a two-level quantum system.

• Unitary Evolution: The evolution is lossless, meaning there is no decoherence or dissipa-

tion.

1.5 Oracle in Continuous-Time Formalism

In the continuous-time framework, the oracle operator is encoded in the Hamiltonian as Ĥoracle :

Ĥoracle = |w⟩⟨w|
This term ensures that the marked state |w⟩ is an eigenstate of Ĥoracle with a distinct eigen-
value (e.g., 1), while all other states have eigenvalue 0. When combined with Ĥdiffusion , the total
Hamiltonian creates constructive interference in the probability amplitude of |w⟩, amplifying its
amplitude over time.

2 Main Differences Between Discrete-Time and Continuous-

Time Implementations
2.1 Evolution Framework
• Discrete-Time Implementation: Grover’s algorithm uses a sequence of discrete
unitary operations (oracle and diffusion operators) applied iteratively to the quantum
state. Each step corresponds to a specific matrix operation.

• Continuous-Time Implementation: The quantum state evolves under a time-

dependent Hamiltonian governed by Schrödinger’s equation. The oracle and diffusion
effects are embedded in the Hamiltonian.

2
2.2 Dynamics
• Discrete: Iterative and periodic, with constructive interference occurring at specific
iterations.

• Continuous: Smooth oscillatory evolution in the state amplitudes driven by the

Hamiltonian.

2.3 Complexity
√
Both implementations achieve O( N ) scaling, but the continuous-time version computes
the total evolution time rather than discrete iterations.

2.4 Implementation
• Discrete: Requires a clocked sequence of gates.

• Continuous: Relies on analog control of the Hamiltonian, which is harder to imple-

ment precisely in physical systems.

3 Differences in Adiabatic Quantum Computing Frame-

work
3.1 Key Idea of Adiabatic Quantum Computing
In adiabatic quantum computing (AQC), the system starts in the ground state of an initial
Hamiltonian Ĥinit . The Hamiltonian is then slowly interpolated to the problem Hamiltonian
Ĥoracle :
   
t t
Ĥ(t) = 1 − Ĥinit + Ĥoracle
T T
The solution is encoded in the ground state of Ĥoracle .

3.2 Differences from Grover’s Hamiltonian Framework

• In AQC, the system must remain in the ground state throughout the evolution, requir-
ing adiabatic (slow) changes to the Hamiltonian.

• Grover’s continuous-time algorithm does not require the system to stay in the ground
state and leverages oscillatory dynamics.

3.3 Efficiency
AQC’s runtime depends on the minimum energy√ gap between the ground state and excited
states, which may scale differently than O( N ).

3
4 Interference Principle in Hamiltonian-Based Grover’s
Algorithm
In Grover’s algorithm, constructive and destructive interference play a critical role in guiding
the quantum system towards the marked state (the solution) by modifying the amplitude
of different states. These interferences emerge from the quantum evolution governed by the
Hamiltonian, which describes the time evolution of the system.
Here’s how constructive and destructive interference arise and guide the state towards
the solution in a Hamiltonian-based derivation:

4.1 1. Hamiltonian-Based Derivation

In the Hamiltonian-based continuous-time version of Grover’s algorithm, the state evolves
according to the time-dependent Schrödinger equation:
d
|ψ(t)⟩ = −iĤ(t) |ψ(t)⟩
dt
Where Ĥ(t) is the Hamiltonian that governs the evolution of the system. In the context
of Grover’s search problem, the Hamiltonian typically has two parts:

• Oracle Hamiltonian: Encodes the problem and flips the phase of the marked state.

• Diffusion Hamiltonian: Amplifies the amplitude of the marked state and decreases
the amplitude of the unmarked states.

The total Hamiltonian can be written as:

Ĥ = Ĥoracle + Ĥdiffusion

4.2 2. Oracle Hamiltonian (Encoding the Search Problem)

The oracle operator Ô marks the correct state by applying a phase flip, i.e., it changes the
sign of the amplitude of the marked state |w⟩ (the solution to the search problem). The
oracle operator acts as follows:
(
−|x⟩ if x = w
Ô|x⟩ =
|x⟩ otherwise
The oracle Hamiltonian encodes the search problem and provides the initial direction of
interference by flipping the phase of the marked state.

4.3 3. Diffusion Hamiltonian (Amplifying the Correct State)

The diffusion operator (also called the inversion about the average operator) amplifies the
amplitude of the marked state and decreases the amplitude of the other states. It acts as
follows:

Ĥdiffusion = 2|ψ0 ⟩⟨ψ0 | − Iˆ

4
 Where |ψ0 ⟩ is the equal superposition state of all possible solutions and Iˆ is the iden-
tity operator. This operator brings the quantum state closer to the marked state through
constructive interference.

4.4 4. Constructive and Destructive Interference

4.4.1 Constructive Interference
Constructive interference occurs when the quantum amplitudes of states are reinforced. In
the context of Grover’s algorithm, the diffusion operator amplifies the amplitude of the
marked state and those states that are close to it in the Hilbert space.

• The oracle flips the phase of the marked state, introducing a negative sign to its
amplitude.

• The diffusion operator then acts to increase the amplitude of the marked state, effec-
tively applying constructive interference.

• The overall process iteratively boosts the probability of measuring the marked state
by reinforcing its amplitude at each iteration.

4.4.2 Destructive Interference

Destructive interference occurs when the quantum amplitudes of states cancel each other
out, effectively diminishing their probability.

• After applying the oracle operator, the states that are far from the solution get a phase
flip, and the diffusion operator works to reduce the probability of these states.

• These non-solution states (incorrect states) undergo destructive interference, as their

amplitudes are diminished by the diffusion operator.

• Through this iterative process, the amplitude of the non-marked states is gradually
reduced, whereas the amplitude of the marked state is amplified.

4.5 5. Guiding the Quantum State Towards the Solution

The key idea behind constructive and destructive interference is that the amplitude of the
marked state grows over time while the amplitude of other states decreases. In the Hamil-
tonian representation of Grover’s algorithm:

• Constructive interference guides the quantum state towards the solution by amplifying
the probability amplitude of the marked state.

• Destructive interference reduces the probability of observing non-solution states by

diminishing their amplitudes.

This process is repeated

√ iteratively. The quantum state is evolved step by step in such
a way that, after O( N ) steps, the probability of measuring the marked state reaches its
maximum value.

5
4.6 6. Mathematical Representation of Interference
After each application of the oracle and diffusion operations, the quantum state evolves in
such a way that:

|ψ(t)⟩ = cos(θ)|marked state⟩ + sin(θ)|unmarked state⟩

Where θ is the angle between the current quantum state and the marked state, and the
probability of measuring the marked state is proportional to cos2 (θ). The amplitude of the
marked state grows over time due to the constructive interference introduced by the diffusion
operator, and the amplitudes of unmarked states decay due to destructive interference.