Quantum computing under real-world constraints: Efficiency of an ensemble quantum algorithm and fighting decoherence by gate design

Quantum algorithms for problems such as integer factorization and database search have spurred great interest in quantum computation. However, the construction of scalable quantum computers has proved to be extremely difficult. Among the challenging requirements that must be fulfilled in the "standard model" of quantum computation, one must be able to initialize qubits into known pure states. In addition, quantum computers must be protected from noise due to unwanted interactions with their environment.; This work is concerned with two theoretical issues motivated by these difficulties. First, I analyze the efficiency of an algorithm for an alternate model for quantum computation that is motivated by bulk nuclear magnetic resonance. In such ensemble systems, the initial state is not a fiducial pure state, but is instead a maximally-mixed density operator. The algorithm that I consider allows one to estimate the free energy of arbitrary spin-1/2 lattice models. To determine the efficiency of this algorithm, I calculate the computation time required to estimate the free energy to bounded error as a function of the lattice size. In the absence of stochastic fluctuations in the measurement output, it is found that the algorithm is efficient. However, evidence is presented that suggests that the algorithm becomes exponentially sensitive to fluctuations as the lattice size increases.; While techniques such as quantum error-correcting codes and decoherence-free subsystem encodings have been devised to mitigate errors due to unwanted environmental couplings, these methods require many additional qubits or complicated encodings. Here, I investigate a simple approach to reduce errors in the quantum search algorithm due to a collective decoherence model. This method takes advantage of the freedom inherent in compiling the search algorithm into fundamental gates. Transition rate calculations and more rigorous quantum master equation simulations are carried out for small-qubit instances to contrast the performance of the original and modified algorithm. It is shown that the expected computational effort can be reduced by 22% for selected five-bit instances of the quantum search algorithm. While this approach does not constitute a general strategy for quantum error-correction, it illustrates the importance of judicious gate design in mitigating decoherence.