Quantum Optimal control with superconducting qubits for quantum information processing.

上世纪80年代，S提出了计算领域的一个新概念：一台通用量子计算机。这种新计算机的工作方式与传统计算机不同，事实上，这种新模式在不同任务上的表现可能会超过传统计算机。量子计算机上的信息是连续的，而在经典计算机上是离散的，这引入了一系列以前不存在的问题。量子比特是量子计算机的最小信息单位，类似于经典计算机中以0或1的形式存在的比特。在实验上，量子比特有许多可以非常不同的实现，我的研究最集中的一个是在Transmon制度下操作的约瑟芬结，这是一种对电荷噪声不敏感的量子比特。为了从这些量子比特中获得任何形式的有用输出，需要进行一些操作。尽管这些运算没有被完美地实现，但存在一个错误率，对于该错误率，使用一些纠错方案，使用量子比特中的一些冗余，计算将是可行的。这种信息的重复可以使量子比特的数量增加一千倍以上！通过获得极高的保真度，即所述操作执行的可靠性的量度，可以减少该开销。实现这些可靠操作的方法是使用微波脉冲。然而，在量子比特参数存在波动或有噪声的情况下，保真度的值会显著下降。需要找到一种策略来对抗这些不良影响：我使用顺序凸编程--一种梯度搜索优化方法--来找到对这些波动和噪声具有健壮性的脉冲。由于对优化初始条件的敏感性，需要使用大量的起点来探索保真度前景，这需要可以同时在多台计算机上运行的代码。在未来，我想看看不同的系统。

In the 80's a new idea in computing was proposed: a universal quantum computer. This new computer would function differently to how a classical computer works, and in fact the new paradigm could outperform the classical computer at different tasks. The information on a quantum computer is continuous while on a classical computer it is discrete and this introduces a whole range of problems that did not exist previously.Qubits are the smallest units of information for a quantum computer, analogous to the bit which exists as either 0 or 1 in a classical computer. Experimentally, qubits have many implementation that can be vastly different, the one that my research is most focussed toward is the Josephon junction operated in the Transmon regime, a qubit that is not sensitive to charge noise. In order to get any sort of useful output from these qubits some operations need to be carried out. Despite these operations not being achieved perfectly there exists an error rate for which computation will be feasible with some error correcting scheme, using some redundancy in the qubit. This repetition of information can increase the number of qubits by a factor greater than a thousand! By obtaining extremely high fidelity, a measure of the reliability of the execution of said operation, this overhead can be reduced. The way these reliable operations are achieved is by using microwave pulses. However, in the presence of fluctuations in the parameters of the qubit or when there is noise, the value for the fidelity drops significantly. A strategy needs to be found to fight those undesirable effects: I use sequential convex programming, a gradient search optimisation method, to find pulses that will be robust to these fluctuations and noise. Due to the sensitivity to the initial condition of the optimisation a large number of starting points need to be used to explore the fidelity landscape this requires code that can be run on multiple computers at once. In the future, I would like to look at different systems.