Optimal control of open quantum systems in strong coupling regimes

Quantum technologies (QTs) have the potential to revolutionise fields like computing, communication, and sensing by leveraging the unique properties of quantum dynamics, enabling information processing beyond classical devices. However, precise control over these fragile quantum systems presents a significant challenge to realising QTs. Even in idealised, isolated systems, rapidly changing the system induces undesirable transitions, disrupting the control protocol and introduces errors. This creates a fundamental trade-off: faster operations increase the likelihood of errors, while slower operations, though more reliable, are impractical for real-world applications. The challenge becomes even more complex in practical quantum devices, which are never fully isolated from their surroundings and are often significantly impacted by environmental fluctuations. These environmental interactions, such as those between a system’s electronic degrees of freedom and the vibrational modes of the surrounding material, lead to decoherence and dissipation which, in-effect, erases quantum phenomena necessary for QTs, thereby degrading the performance of a device. Solid-state quantum devices, for instance, face significant challenges in maintaining a quantum state under these conditions, limiting their scalability. Therefore, to ensure reliable operation, accounting for these environmental influences is essential when designing control protocols for real-world quantum systems. This is particularly challenging, as the complexity of environmental interactions often prevents their effects from being accurately captured using conventional methods which assume weak environmental coupling., This project seeks to overcome these challenges by integrating cutting-edge methods from open quantum systems theory with machine learning tools. Open quantum systems theory allows us to model the interactions between quantum systems and their environment, providing insights into how these interactions degrade performance. Central to our approach is integrating tensor network methods—efficient numerical techniques for simulating quantum systems with strong environmental coupling—with machine learning tools, such as Bayesian neural networks. This hybrid framework will enable the design of optimal control protocols for complex systems, significantly reducing computation time compared to traditional methods. Ultimately, this project aims to develop control strategies that effectively counteract environmental interactions while enhancing the speed of quantum operations. A key focus will be on developing Shortcuts to Adiabaticity (STA) for open quantum systems strongly coupled to their environment. In isolated systems, STA protocols cancel the unwanted transitions that occur when control fields are applied rapidly, reducing the trade-off between fast quantum control and the accumulation of errors. However, developing such protocols for strongly coupled open systems remains an open challenge. The theoretical framework developed in this project provides a novel approach to addressing this challenge with implications for solid-state QTs, such as quantum dots and defect centres in crystaline materials. By collaborating with experimental partners, the developed methods will account for real-world constraints and complexities in experimental setups, ensuring the practicality of the resulting protocols. This project will provide a flexible theoretical framework capable of addressing key challenges in the control of realistic quantum systems. The aim is to make QTs more robust and scalable, bringing us closer to their widespread application in various industries, from secure communication to computation, helping to drive the next generation of technological innovation.