A key function of a quantum internet is the generation of entangled links between devices. The quality of these links decays over time due to interaction with the outside world. Various protocols exist for generating these links. The main trade-off to be considered when choosing a protocol is between the chance of producing the link and the quality of the generated link. Selecting the optimal protocol is a complex and computationally intensive task, especially as the scale of the problem increases. While analytical solutions are feasible for small-scale systems, they become impractical for larger networks due to their computational demands. In addition, the rate at which the generated links decay is usually assumed to be static, when in reality it typically fluctuates over time in a process called parameter drift. In this paper, we evaluate the effectiveness of reinforcement learning approaches to optimizing protocol selection for different models of this problem. Models are provided for a static rate of decay, for a drifting rate of decay, and for the phenomenon of induced decoherence. Different reinforcement learning agents are trained on all of these models, and the results are compared. The criterion for effectiveness is the speed at which the requested number of entangled links can be generated. A negligible effect on performance is detected, showing that models are able to adapt to the different sources of instability studied. We also provide methods to model this problem and identify promising directions for future research.

Many quantum internet applications need access to multiple entangled links existing simultaneously. This requires the generation of multiple entangled links within a time window. Using single entangled link generation protocol, we model this task as a Markov decision process and propose a heuristic-based policy to generate them. This policy chooses different configuration parameters depending on the number of links in the register. Entanglement purification is also incorporated into this heuristic in different ways. To compare, we choose two baseline policies including a previously studied fixed configuration parameter policy. We show that this algorithm completes the generation up to six times faster than the baselines in our simulations and that using entanglement purification can further improve its performance.

Entangled links can be seen as a connection between two parties that persists remotely, but whose quality (fidelity) decreases over time. Many quantum applications rely on having a certain number of simultaneously active entangled links for their execution. Every application has a characteristic threshold fidelity - the minimum fidelity a link should have for it to be useful. There are several link generation protocols; a protocol i generates with probability P_i a link with initial fidelity F_i. Until now, research has focused on generating all links with the same protocol. Therefore, the aim of this research is to use multiple generation protocols in order to minimize the expected time needed to reach the desired number of links, and to analyze how this time changes with regard to the threshold fidelity of the target protocol. We present how to reach the minimum possible expected time by modelling the problem as a Markov decision process and determining the optimal sequence of protocols to be used. Then, we observe that our approach generally performs several orders of magnitude better than previous research in which all links were generated with the same protocol.

Quantum networks allow quantum processors to communicate over large distances. These networks often require simultaneously existing multiple entangled pairs of quantum bits (entangled links) as a fundamental resource for communication. Link generation is a sequential and probabilistic process, and successfully generated links are stored in a quantum memory. Links in memory are subject to noise that causes their quality to decay and become unusable. This paper uses reinforcement learning (RL) to investigate dynamic tuning of the entanglement generation protocol to minimise the time to generate multiple links. By comparing a fixed number of actions to a continuous action space, we analyse the importance of finer-grained tunings of the protocol. This is tested in simulated near-term and medium-term network abstractions. The results show that protocol tuning significantly reduces the mean time to generate entangled links, with finer tuning providing greater benefits up to a point. Furthermore, a heuristic is derived from the RL policies which matches and exceeds their performance. Future work can explore more advanced reinforcement learning algorithms to find better policies, as well as using different noise models to make more generally applicable policies.

A quantum network allows us to connect quantum information processors to achieve capabilities that are not possible using classical computation. Quantum network protocols typically require several entangled states available simultaneously. Previously, an entanglement generation process was analysed where, at each time step, we generate an entangled state with success probability p. Here, we consider adaptive entangled state generation with more flexibility. At each time step, our process chooses a protocol (p_i, F_i) from a discrete number of entanglement generation protocols. An entangled state is generated successfully with probability p_i, and its fidelity F_i defines how close the entangled state is to an ideal Bell state. The new state is subject to depolarising noise in the quantum memory. Because of the memory noise, states are discarded after a certain number of time steps t_i when they are no longer useful to our application. We model our process as a Markov decision process and derive a policy π to generate n entangled states with minimal expected time Eπ[τ]. We analyse the offered improvement of the optimal policy of our adaptive entanglement generation process over the previously studied static process. We conclude that this improvement becomes more significant as the required number of links in memory increases.

Entanglement can be used as a resource to support a wide range of quantum applications. However, the scarcity of efficient entanglement distribution protocols poses a significant challenge for the deployment of large-scale quantum networks. Losses in the media prevent the direct transmission of quantum states over large distances, but the use of quantum repeaters presents a possible alternative for long-distance quantum communication. Here, we focus on homogeneous quantum repeater chains and provide some guidelines based on heuristic methods that allow the design of entanglement swapping policies. For instance, delaying simultaneous swaps on adjacent nodes can reduce the probability of losing entanglement. Whereas previous work mainly focused on chains with few nodes only, we present three different policies that are easy to implement and scalable to longer chains. We evaluate these policies using Monte Carlo simulations, comparing their performance to the well-known swap-asap policy. When classical communication time is neglected, our policies provide lower delivery time than swap-asap for probabilistic swaps and large entanglement generation probability. When classical communication time is large, only one of our policies is in most cases faster than swap-asap for both probabilistic and deterministic swaps.