grover. part 2

Post on 03-Jan-2016

48 Views

Category:

Documents

4 Downloads

Preview:

Click to see full reader

DESCRIPTION

Grover. Part 2. Components of Grover Loop. The Oracle -- O The Hadamard Transforms -- H The Zero State Phase Shift -- Z. O is an Oracle. H is Hadamards. Z is Zero State Phase Shift. H is Hadamards. Grover Iterate. Inputs oracle. This is action of quantum oracle. - PowerPoint PPT Presentation

TRANSCRIPT

Grover. Part 2

Components of Grover Loop

• The Oracle -- O

• The Hadamard Transforms -- H

• The Zero State Phase Shift -- Z

O is an Oracle

H is Hadamards

H is Hadamards

Z is Zero State Phase Shift

Grover Iterate

Inputs oracle

We need to initialize in a superposed state

This is action of quantum oracle

This is a typical way how oracle operates

This is a typical way how oracle operation is described

Encodes input combination with changed sign in a superposition of all

Role of Oracle

• We want to encode input combination with changed sign in a superposition of all states.

• This is done by Oracle together with Hadamards.

• We need a circuit to distinguish somehow globally good and bad states.

Vector of Hadamards

Notation Reminder

a Control with value a=1

a Control with value a=0

a

equivalent

This is value of oracle bit

Flips the data phase

All information of oracle is in the phase but how to read it?

This is just an example of a single minterm, but can be any function

Zero State Phase Shift Circuit

Flips the oracle bit when all bits are zero

Rewriting matrix Z to Dirac notation, you can change phase globally

This is state of all zeros

2 0 0 0

0 0 0 0

0 0 0 0

0 0 0 0

1 0 0 0

0 -1 0 0

0 0 -1 0

0 0 0 -1

-1 0 0 0

0 -1 0 0

0 0 -1 0

0 0 0 -1

+ =

With accuracy to phase

In each G

This is a global view of Grover. Repeatitions of G

Here you have all components of Grover’s loop

Generality

• Observe that a problem is described only by Oracle.

• So by changing the Oracle you can have your own quantum algorithm.

• You can still improve the Grover loop for particular special cases

proof

Here we explain in detail what

happens inside G. This can be

generalized to G-like circuits

Grover iterate has two tasks: (1) invert the solution states and (2) invert all states about the mean

Here we prove that |> < | used inside HZH calculates the mean

a Vector of mean values

Will be explained in next slide

Explanation of the first part of Grover iterate formula

This proof is easy and it only uses formalisms that we already know.

(( ))

(( ))

From previous slide

What does it mean invert all states about the mean?

For every bit

Amplitudes of bits after Hadamard

Positive or negative amplitudes in other explanations

All possible states

Amplitudes of bits after one stage of G

This value based on previous slide

This slides explains the basic mechanism of the Grover-like algorithms

You can verify it also in simulation

Additional Additional ExerciseExercise

This is a lot calculations, requires matrix multiplication

Here we calculate analytically when to stop

The equations taken from the previous slides “Grover Iterate”

For marked state

For unmarked state

We found k from these equations

recursionWe want to find how many times to iterate

But you can do better if you have knowledge, for instance the upper bound of chromatic number in graph coloring

Grover search example.

• Here is an example of Grover search for n = 3 qubits, where N = 2n =8. – We omit reference to qubit n+1, which is in

state 1 /√2 (|0>−|1>i) and does not change.

• The dimension of the unitary operators for this example is thus 2n = 8 also.)

• (Remember that numbering starts with 0 and ends with 7, so that the -1 here is in the slot for |5>.)

• This matrix reverses the sign on state |5>, and leaves the other states unchanged.

•Suppose the unknown number is |a> = |5>. •The matrix or black box oracle Ufa is

oracle

•The Walsh matrix W8 is

Now we use normalization

The matrix −Uf0 is

This matrix changes the sign on all states except |0>.

Finally, we have the repeated step RsRa in the

Grover algorithm:

oracle

shift

hadamards

After second rotation we get

Summary and our work

When you know anything about the problem (symmetry, observation, bounds, function within some classification class) you can design a better Grover like algorithm but for your data only.

This is enough in real life like CAD or Image Processing, since data are always specific, not the worst case data as in Mathematic proofs

Problem for students

• Build the Grover algorithm for ternary quantum logic.

• First you need to generalize Hadamard transform to Chrestenson transform.

• Next you need to have some kind of ternary reversible gates to build oracle.

• The same gates will be used for Zero State Phase Shift circuit.

top related