This paper investigates the subcycling of particle orbits in variational, geometric particle-in-cell methods, addressing the Vlasov-Maxwell system in magnetized plasmas. The purpose of subcycling is to allow different time steps for different particle species and, ideally, time steps longer than the electron gyroperiod for the global field solves while sampling the local cyclotron orbits accurately. The considered algorithms retain the electromagnetic gauge invariance of the discrete action, guaranteeing a local charge conservation law, while the variational approach provides a bounded long-time energy behavior.

This paper proposes a novel numerical integrator for modeling multispecies Coulomb collisions in kinetic plasmas. The proposed scheme provides an energy-, momentum-, and positivity-preserving particle discretization of the nonlinear Landau collision operator, extending the works of Carrillo et al. [J. Comput. Phys. 7, 100066 (2020)] and Hirvijoki [Plasma Phys. Controlled Fusion 63, 044003 (2021)]. The discrete-time conservation properties are analyzed both algebraically and numerically, and an efficient, graphics processing unit-parallelized implementation is validated against inhomogeneous temperature relaxation, isotropization, and thermalization examples. The results agree with analytical estimates, confirming the method capable of reproducing physics.

Recently, a new approach to gyrokinetics, invariant under electromagnetic gauge transformations, was developed. The gyrocenter equations of motion are now expressed in terms of the perturbed fields instead of the potentials, in a form suitable for numerical simulations and analytic studies. In this paper, we verify that the long-wavelength limit, i.e., the drift-kinetic limit of the new gyrokinetic theory, is in line with existing work, providing a solid foundation for simulations. We compute the dispersion relation of the new drift-kinetic theory in slab geometry and find agreement with a long-wavelength limit of the full Vlasov-Maxwell model.

This paper presents a novel scheme to improve the statistics of simulated fast-ion loss signals and power loads to plasma-facing components in fusion devices. With the so-called Backward Monte Carlo method, the probabilities of marker particles reaching a chosen target surface can be approximately traced from the target back into the plasma. Utilizing the probabilities as {\it a priori} information for the well-established Forward Monte Carlo method, statistics in fast-ion simulations are significantly improved. For testing purposes, the scheme has been implemented to the ASCOT suite of codes and applied to a realistic ASDEX Upgrade configuration of beam-ion distributions.