QSharp-Playground
Playing around with Q#
Trying to learn some of the basic concepts of quantum computing with Q#. Q# is Microsoft's domain-specific programming language used for expressing quantum algorithms. It is part of the Microsoft Quantum Development Kit - which is an early preview version.
You will find in this repository:
- an implementation of a quantum 4-bit adder
- a oracle function using this 4-bit adder
- leverage this oracle function by applying Grover's algorithm in order to search for two summands which give a specified sum
4-qubit adder
The implementation of the 4-bit adder follows the approach detailed here. This Conventional Quantum Plain Adder takes two 4-qubit inputs and produces a 5-qubit output. In addition, we need to have 4 ancillary qubits.
The prototype of the adder is:
operation FourBitAdder(a1: Qubit,a2: Qubit,a3: Qubit,a4: Qubit,b1: Qubit,b2: Qubit,b3: Qubit,b4: Qubit,carry:Qubit, tempQubits:Qubit[]):()
Notes:
- The result is placed in the registers b1, b2, b3, b4, carry.
- Quantum operations have to be reversible, which this gate ensures.
- The 4 ancillary registers (tempQubits) have to be in pure state |0> on input, and they are |0> on output.
- The carry-qubit has to be in state |0> on input.
4-qubit adder in action
The command-line program "4BitAdderAndGrover.exe" allows to call this quantum 4-bit-adder with the inputs in a mixed state, and in a sense we calculate the sum for all possible inputs at once - and if we do a measurement on the result, we get one of those possibilities.
Here is an example:
>.\4BitAdderAndGrover.exe -o TestAdderWithEntangledInput -r 10
Executing the 4-bit adder with entangled inputs
20 = 7 + 13
19 = 11 + 8
26 = 14 + 12
5 = 1 + 4
23 = 12 + 11
11 = 4 + 7
11 = 5 + 6
8 = 7 + 1
13 = 11 + 2
16 = 5 + 11
*** Everything OK. ***
Of course - this is quite useless. It is mind-blowing that the quantum computer is able to do the computation for all possible inputs at once in one step, but what's the use if we are only able to get one of those results - and we are not even able to control which one. Or... can we?
Grover's Algorithm
It seems to turn out, that we actually can control (to some extent) which of the results we get when measuring. That's what Grover's Search Algorithm is about (I found this paper very instructive). The ingredient to Grover's algorithm is an oracle function which has the task of
- bring the input/output qubits in superposition
- determine whether the input qubits give the result we are looking for
- if so, set the flag-qubit to |1>
This allows us to specify a result, and then determine the input which gives the result. So, we can solve the question "for which input do I get this result?".
We need to repeat Grover's algorithm times in order to get maximum likelihood of measuring the desired result (where the flag-qubit is |1>) - where n is the number of input qubits and k is the multiplicity of "good inputs". The latter is unknown in general - which then goes beyond the basic application of Grover.
This example determines which summands give a sum of 29 (and repeat this calculation 3 times):
> .\4BitAdderAndGrover.exe --operation FindSummands -e 29 -g 9 -r 3
Finding the summands which give the specified sum (29), using 9 Grover-iterations.
1: a=15 b=14 ; 751 0000'1'0'1110'1111
2: a=15 b=14 ; 751 0000'1'0'1110'1111
3: a=14 b=15 ; 766 0000'1'0'1111'1110
Grover-Iterations 9: 3 of 3 had the desired result.
We had to iterate 9 times because is approximately 9 (and there are two possible sets of summands). If we use an inappropriate number of iterations, we are more likely to measure a "bad state":
> .\4BitAdderAndGrover.exe --operation FindSummands -e 29 -g 3 -r 5
Finding the summands which give the specified sum (29), using 3 Grover-iterations.
1: a=15 b=14 ; 751 0000'1'0'1110'1111
2: not successful
3: a=14 b=15 ; 766 0000'1'0'1111'1110
4: a=15 b=14 ; 751 0000'1'0'1110'1111
5: not successful
Grover-Iterations 3: 3 of 5 had the desired result.