The last two decades have seen an enormous increase in the computational power of digital computers. This was due to the rapid technical development in manufacturing processes and controlling semiconducting structures on submicron scale. Concurrently, the electric circuits have encountered the first signs of the realm of quantum mechanics. Those effects may induce noise and thus they are typically considered harmful. However, the manipulation of the coherent quantum states might turn out be the basis of powerful computers – quantum computers. There, the computation is encoded into the unitary time evolution of a quantum mechanical state vector. Eventually, quantum mechanics could enable one, for example, to read secret electronic messages which are encrypted by the widely employed RSA cryptosystem – a task which is extremely laborious for the current digital computers. This thesis presents a theoretical study of the coherent manipulations of pure quantum states in a quantum register, that is, quantum algorithms. An implementation of a quantum algorithm involves the initialization of the input state and its manipulation with quantum gates followed by the measurements. The physical implementation of each gate requires that it is decomposed into low-level gates whose physical realizations are explicitly known. Here, the problem is examined from two directions. Firstly, the numerical optimization scheme for controlling time-evolution of a closed quantum system is discussed. This yields a method for implementing quantum gates acting on up to three quantum bits, qubits. The approach is independent of the physical realization of the quantum computer, but it is considered explicitly for a proposed inductively coupled Josephson charge qubit register. Secondly, the techniques of numerical matrix computation are utilized to find a general method for decomposing an arbitrary n-qubit gate into a sequence of elementary gates, which act on one or two qubits. The results of this thesis help to improve the implementation of quantum algorithms. The quantum circuit construction developed in the thesis is the first one to achieve the asymptotically minimal complexity in the number of elementary gates. In context of acceleration of quantum algorithms we present a gate-level study of Shor's algorithm and show how to accelerate the algorithm by merging several elementary gates into multiqubit gates. Finally, the requirements set by the resulting gate array are compared to the properties of superconducting qubits. This allows us to discuss the feasibility of the Josephson charge qubit register, for instance, as hardware for breaking the RSA cryptosystem.
The last two decades have seen an enormous increase in the computational power of digital computers. This was due to the rapid technical development in manufacturing processes and controlling semiconducting structures on submicron scale. Concurrently, the electric circuits have encountered the first signs of the realm of quantum mechanics. Those effects may induce noise and thus they are typically considered harmful. However, the manipulation of the coherent quantum states might turn out be the basis of powerful computers – quantum computers. There, the computation is encoded into the unitary time evolution of a quantum mechanical state vector. Eventually, quantum mechanics could enable one, for example, to read secret electronic messages which are encrypted by the widely employed RSA cryptosystem – a task which is extremely laborious for the current digital computers. This thesis presents a theoretical study of the coherent manipulations of pure quantum states in a quantum register, that is, quantum algorithms. An implementation of a quantum algorithm involves the initialization of the input state and its manipulation with quantum gates followed by the measurements. The physical implementation of each gate requires that it is decomposed into low-level gates whose physical realizations are explicitly known. Here, the problem is examined from two directions. Firstly, the numerical optimization scheme for controlling time-evolution of a closed quantum system is discussed. This yields a method for implementing quantum gates acting on up to three quantum bits, qubits. The approach is independent of the physical realization of the quantum computer, but it is considered explicitly for a proposed inductively coupled Josephson charge qubit register. Secondly, the techniques of numerical matrix computation are utilized to find a general method for decomposing an arbitrary n-qubit gate into a sequence of elementary gates, which act on one or two qubits. The results of this thesis help to improve the implementation of quantum algorithms. The quantum circuit construction developed in the thesis is the first one to achieve the asymptotically minimal complexity in the number of elementary gates. In context of acceleration of quantum algorithms we present a gate-level study of Shor's algorithm and show how to accelerate the algorithm by merging several elementary gates into multiqubit gates. Finally, the requirements set by the resulting gate array are compared to the properties of superconducting qubits. This allows us to discuss the feasibility of the Josephson charge qubit register, for instance, as hardware for breaking the RSA cryptosystem.
Quantum computing and quantum information science are two recently discovered and rapidly growing fields of physics that show substantial promise in providing new and valuable technologies in the foreseeable future. Large-scale quantum computers, if ever realized experimentally, are likely to outperform their classical counterparts in a number of important computational tasks, the most important of which may be the accurate simulation of many-body quantum systems such as the ones encountered in physics, chemistry and life sciences. This thesis investigates the problem of controlling quantum systems for the purpose of performing quantum information processing tasks. The problem is approached from a theoretical and simulational viewpoint. The work contained here encompasses a range of levels of abstraction. Firstly, we discuss the decomposition of abstract multiqubit logic gates into sequences of simple elementary gates. Secondly, we study the local commutational properties of two-qubit gates using local gate invariants. Thirdly, we develop methods for the physical implementation of the elementary gates through the control of specific quantum systems, possibly in the presence of noise and decoherence. We present a new, almost optimal n-qubit gate decomposition based on the cosine-sine decomposition, which utilizes a likewise new intermediate quantum circuit structure we call a uniformly controlled gate. We then show how they can be used in constructing a general state transformation circuit. Both of the resulting circuits can be efficiently implemented using nearest-neighbor gates which makes their physical realization simpler. A local gate invariant is introduced which can be used to assess the suitability of two-qubit gates for serving as the entangling gate in elementary gate libraries. Finally, we develop numerical optimization methods for finding near-optimal control sequences for generating oneand two-qubit gates, both in closed quantum systems and in the presence of Markovian noise.
The concept of a quantum computer was invented in the beginning of the 1980s as a quantum generalization of the reversible classical computer. After discoveries in the middle of the 1990s of quantum algorithms, which wouldsolve some problems considered intractable for classical computers, the field of quantum computing has developed rapidly. A wide range of quantum systems are investigated for their possible ability to implement quantum algorithms in practice. The most promising solid state implementations are superconducting electrical circuits based on the Josephson effect, and the Coulomb blockade effect. Superconducting circuits are macroscopic quantum systems which can be fabricated using standard lithography technologies. Superconducting qubits - basic building blocks of a quantum computer - have been developed at several research laboratories, among others the MC2 here at Chalmers. This thesis presents a theoretical investigation of controllable coupling of superconducting qubits based on the single Cooper-pair box. Two coupling schemes are investigated in detail: current-controlled coupling via large Josephson junctions, and coupling via a superconducting stripline cavity. Both coupling designs are scalable and suitable for the charge regime of the single Cooper-pair box, as well as for the charge-phase regime where the qubit is better protected from environmental noise. Quantum gate protocols which are relevant for these physical couplings are discussed. The investigations show that, for both considered designs, the simplest two-qubit gate is the conditional phase-shift gate. This gate is universal for quantum computing, and it can create maximallyentangled qubit pairs in only one run. The investigation combines methods from electrical engineering, quantum mechanics and quantum computer science.
A quantum computer is expected to solve some problems far faster than today's computers and may be a promising device with low power consumption.This computer is based on the physical theory of quantum mechanics.In the theory of quantum mechanics, quantum states and evolutions of quantum states are represented by vectors in a Hilbert space and unitary operators, respectively.A unitary matrix can then be regarded as an algorithm in the conventional computer.Since it is generally difficult to develop an arbitrary unitary matrix in a quantum system, quantum computation is carried out by combining elementary operations such as one-or two-qubit operations, where a qubit is the smallest unit of information in quantum computation.A unitary matrix is said to be performed effectively if there exists a quantum circuit composed of a polynomial number of elementary gates.Therefore, translating a unitary matrix into an efficient sequence of elementary gates is a fundamental problem in designing quantum circuits.In this dissertation, the problem involved in synthesizing minimal quantum circuits for carrying out any (large) quantum operation is described.The author considers two quantum systems, namely, the two-level quantum system and the d-level quantum system, where d is any integer greater than two.Here, the two-level quantum system can be regarded as a quantum mechanical analogue of conventional computation, and the d-level quantum system can be regarded as a quantum mechanical analogue of conventional multilevel logic.
Any unitary transformation of quantum computational networks is explicitly decomposed, in an exact and unified form, into a sequence of a limited number of one-qubit quantum gates and the two-qubit diagonal gates that have diagonal unitary representation in usual computational basis. This decomposition may be simplified greatly with the help of the properties of the finite-dimensional multiple-quantum operator algebra spaces of a quantum system and the specific properties of a given quantum algorithm. As elementary building blocks of quantum computation, the two-qubit diagonal gates and one-qubit gates may be constructed physically with one- and two-body interactions in a two-state quantum system and hence could be conveniently realized experimentally. The present work will be helpful for implementing generally any N-qubit quantum computation in those feasible two-state quantum systems and determining conveniently the time evolution of these systems in course of quantum computation.
Ever since Peter Shor's ground-breaking discovery in 1994 of an algorithm capable of factoring large integers on a quantum-mechanical computer exponentially faster than using any known classical method, research on quantum computing has boomed. Quantum information – a unique mixture of computer science, physics and mathematics – has developed into a new branch of information theory. On the experimental side, physicists from many different disciplines including atomic, solid-state and low-temperature physics, as well as optics, are striving today towards a practical quantum computer. All the candidate quantum bit (qubit) technologies have one thing in common: They rely on the controlled time-evolution of a closed quantum system, a seemingly paradoxical task.\n\nIn this Thesis the temporal control of quantum systems is studied. The topics included can be divided into two according to the type of temporal evolution; geometrical or dynamical. Geometrical realization-independent methods for quantum computing are studied first. Then the study is extended into dynamical quantum computing and the so-called Josephson charge-qubit register is considered as a test bench. Finally, a spin-off application of the geometrical evolution of a Josephson junction system is studied, i.e. Cooper pair pumping. A novel Cooper pair pump, the Cooper pair "sluice", is introduced.\n\nThe work on quantum computing reported in this Thesis is theoretical while the Cooper pair "sluice" is studied both theoretically and experimentally. Numerical simulations, both sequential and parallel, are used extensively throughout the Thesis. The experiments were carried out under cryogenic mK conditions and the sample fabrication was done using e-beam nanolithography.\n\nBecause the execution time of a quantum algorithm is always limited by the inevitable process of decoherence, it is important to utilize any measure available for accelerating quantum computations. It is found that practical quantum algorithms could greatly benefit from classical computer-aided optimization. Moreover, it is found that even a modest demonstrator of a full quantum algorithm using Josephson charge qubits is just barely realizable within present-day coherence times. However, the experimental part of this Thesis shows clear evidence of the functioning of the "sluice". While the worldwide effort of improving the coherence properties of qubits is underway, the "sluice" could well find practical use, e.g., in metrology in the foreseeable future.
Simulating quantum mechanical evolutions in general is difficult on classical computers because the size of the Hilbert space grows exponentially with the number of particles involved in the simulations. However, a quantum computer, which obeys the laws of quantum mechanics, is believed to perform such simulations more efficiently. The time evolution operator, a unitary matrix, of a quantum system acts as a quantum gate changing the state of the system in time. Therefore, any computation on quantum computers can be interpreted as a unitary transformation on quantum state vectors. Generally, quantum gates are sufficiently simple operations so as to avoid difficulty in the physical implementation. The implementation of a general given computation represented by a unitary matrix on quantum computers requires finding an array of elementary quantum gates describing the desired computation. This describes a matrix decomposition problem known as the quantum circuit design problem. In this dissertation, we address the following questions concerning quantum circuit designs: How hard is it to find suitable quantum circuits for a given computation? What are the possible circuit design methods? How do we apply these to simulations of quantum problems, in particular quantum chemistry problems? To answer these questions, we categorize methods for designing quantum circuits as evolutionary algorithms and deterministic methods. We apply our developed group leaders optimization algorithm, an evolutionary algorithm, to find quantum circuits for known quantum algorithms and the simulation of molecular Hamiltonians. Then, we present a deterministic circuit design approach to produce universal circuits, which can be programmable. Moreover, we show how to use these programmable circuit designs within the quantum phase estimation algorithm to find eigenvalues of nonunitary matrices and study resonances in quantum systems. Finally, the quantum phase estimation algorithm is employed to efficiently solve ranking problems: in particular, multiple network alignment is considered to find the similarities between protein-protein interaction networks and molecular networks.