Professional Documents
Culture Documents
Abstract
Using quantum computing, it is possible to design algorithms for computers that handle
problems in ways that were previously unimaginable. The uncertain nature of a particle in a
magnetic field allows for us to handle multiple data sets simultaneously. The string of quantum
bits can act as all values that a classical computer would be able to represent with the same
algorithms can be computed in more reasonable time. Solving complex algorithms of this
nature would change the very nature of computer science we have established. All research is
polynomial time indicates that the expected runtime can be reduced to a polynomial equation
based on the size of the problems inputs (Fortnow 2009). Non-polynomial problems are
Christian Wesseler
inversely unable to be solved in polynomial time, and are broken down into categories NP, NP-
Hard and NP-Complete. Problems defined as NP have an answer that can be verified in
polynomial time. Given an array of size n, a quicksort algorithm can sort the array in O(n*log(n))
time, but the solution can be verified in linear time. In this regard, algorithm analysis becomes
more heavily based on proof of concept than algorithm design. Using proofs, it is possible to
show that problems in NP can be transformed into similarly difficult or harder problems.
Algorithms with this property are classified as NP-Hard, a set of problems that can represent
any NP problem with a polynomial transformation time. If a problem can represent all problems
within NP, and is verifiable in polynomial time, it is said to be NP-Complete. Defining the time
proves matching solutions for every other NP- Graph of an Unsolved Clique Problem (Skipperkongen)
Complete problem. Due to their non-polynomial nature, many researchers have tried to find a
Christian Wesseler
way to prove that NP is the same as P, and that a polynomial time solution exists for problems
in this category. To better explain the significance of proving the two to be the same, it is the
idea of removing the need to search for the next component of the answer in a search problem.
Many problems in NP-Complete are difficult search problems with many potential solutions to
search through. The best example is the maximum clique problem which, given a graph with V
nodes and E edges, must find the largest collection of nodes with every member sharing an
edge with every other member. Each addition to the collection requires one more check than
the previous node. Each collection must be compared with every other collection to see if more
nodes might have fit if specific nodes were excluded. Simplifying the list of potential solutions,
advantage of the unrealistic computation times for NP-Complete problems, much of modern
cryptology hides data encoded as one of these problems, where it is quick to verify the solution
but nearly impossible to compute it. In this sense, solving NP-Complete problems would allow
for users to programmatically break any public key cryptography (Fortnow 2009). Solving NP
problems in polynomial time would allow for an optimization of the world in more ways we
could imagine. To best state the importance of efficiently solving an NP problem, what we will
gain from P = NP will make the whole Internet look like a footnote in history (Fortnow 2009).
Since the P vs NP problem first appeared in Gdels letter to Von Neumann in 1956,
mathematicians have worked tirelessly to create evidence as to the whether P is equal to NP, or
Christian Wesseler
if the two remain distinctly separate (Gdel 1956). In attempting to find a definitive answer, the
only concise proof regarding P vs NP is that no relativizable proof can settle the P versus NP
problem in either direction (Fortnow 2009). It is important to break away from the ideology of
merging the two categories, and instead find a way to transform NP-Complete into an a
problem more closely resembling an equivalent classical polynomial time problem. A solution
would require a reduction in the number of steps involved for finding the optimal solution given
some arbitrarily sized collection of data. The best way to trivialize searching for a solution in this
case is to reduce the impact from the number of inputs on the computation time for the
algorithm. In this case, it would be better to compute all potential solutions simultaneously; a
perform bitwise functions, or provide input to other components of the computer. In this
regard, classical computers are limited in their throughput by how quickly they perform the
operation and load the next instruction to process. Sequential data accesses and computations
Christian Wesseler
are the very concept of the original Turing machine of which the P vs NP problem is based.
computations do not depend on data resulting from previous instructions. Where the classical
bit can encode either a one or a zero, the quantum equivalent, known as a qubit, can represent
both simultaneously through either spinning up or spinning down. This state, known as a
Using qubits, computations can act on all possible states representable by a string of classical
bits. Superposition is achieved by placing a particle in a suspended state, within a magnetic field
so that it exists between both possible spins (Beth and Leuchs 2005). When measured, the spin
is determined to be either up or down, representing a classical bit. The indeterminate state the
bit is in until measured is what gave rise to the infamous Schrdingers Cat thought experiment,
where the fate of the cat is unknown until observed and the box contains both possible
outcomes. The collapse of the possibilities for the particles spin when observed is the basis for
converting qubits into classical bits to decode the information stored. The limitation of
quantum computing is the amount of isolation necessary to prevent the qubit from being
observed before the computation is completed. Any interaction with environmental variables
could prematurely collapse the wave function and prevent further calculations.
In a typical computer system, the processor is unable to function without input from
instructions and data being passed in from memory. The data is stored in registers, passed into
Christian Wesseler
the processor, and performs the instruction pointed to by the program counter. In quantum
computing, each stage in a classical computer would destroy the superposition of the bits in
use. The processor would need to analyze the bits to create the correct output, and the register
would have to check what values to pass to the processor. In order to compute data similar to
the way a classical computer would, the qubits must be isolated from the rest of the system,
and any interactions with them must be independent of the particles spin as to avoid
measuring it. To avoid accessing the data, the qubits would need to act like a cache on a
processor, holding the values locally without external references. The entire dataset for the
problem would need to be encoded as a set of qubits of sufficient size so that no memory
accesses are needed to store the entirety of the NP-Complete problem. Considering the cost of
cache memory deters manufacturers from eliminating random access memory from classical
unrealistic vision for the near future (Bryant and O'Hallaron 2013).
Given the amount of work in creating a system that could reasonably isolate enough particles
to recreate an NP-Complete problem, the question remains what would be the improvement
for runtime performance? Reusing the example of the clique problem, where V represents the
number of vertices present in the graph, we can do a proof by induction to find that every
clique problem has a collection of subgraphs twice the size of a clique problem with one less
vertex. This is further reinforced if imagining all subgraphs of V-1 size alongside each similar
Christian Wesseler
subgraph containing the new vertex. We know that the solution will be one of the 2V sets of
states. Similarly, many sets within the collection are unsuitable to be considered a clique. To
remove an incompatible set from the collection of possibilities, two active nodes in the set
need to be found without an edge connecting them. Eliminating the need to analyze each set
separately, a new collection of sets can be computed where each set is a completed clique
subgraph. Reducing the collection through quantum computing would eliminate the 2V
coefficient from the algorithms classical complexity and vastly simplify the algorithms
quantum complexity. Each active node would then only need to be compared with every other
active node in the current set to confirm or reject the clique, and the algorithms runtime would
only take as long as the time to verify the largest clique in the graph has no missing edges. Once
a collection has removed all invalid cliques, potential solutions can be parsed for the greatest
number of vertices for the largest clique, and the collection can be sorted like an array for an
ordered list of clique sizes. Through the transformation of the clique problem into a quantum
transformation would not prove polynomial computation time in a classical setting equal to
non-polynomial computation time, but rather a function that converts the problem into a
Conclusion
Using the superposition property of particles, it is possible to develop a quantum computer that
can efficiently solve clique problems. Using the transformations of NP problems into NP-
Christian Wesseler
Complete problems, a quantum computer would be able to efficiently calculate any other
problem in NP by converting it into the clique problem. Once a solution has been calculated, it
can be used to answer the original problem provided as an input. In this regard, there is
potential for quantum computers to work alongside classical computers to convert, compute,
and analyze problems we deemed too complex for Turing machines nearly sixty years ago.
Christian Wesseler
References:
2. Cook, Stephen. "The P versus NP problem." The millennium prize problems (2006): 87-
104.
3. "The Gdel Letter." Gdel's Lost Letter and P=NP. N.p., 12 Feb. 2009. Web. 01 May
2017.
4. McMahon, David. Quantum computing explained. John Wiley & Sons, 2007.
May 2017.
6. Kostas, Author. "Month: November 2010." Skipperkongen. N.p., n.d. Web. 01 May 2017.
7. Beth, Thomas, and Gerd Leuchs. Quantum Information Processing. Weinheim: Wiley-