Speaker
Description
We present a Hamiltonian-based variational Monte Carlo algorithm for (2+1)-dimensional lattice QED with Wilson fermions. The ground state of the model is approximated by a gauge-invariant variational ansatz. The optimal variational parameters are obtained by minimizing the energy expectation value, which can be efficiently evaluated via Monte Carlo sampling for any given parameter set. Unlike most Hamiltonian approaches, we represent the gauge degrees of freedom in a continuous, infinite-dimensional basis, thus avoiding any truncation. The fermionic sector is modeled as a Gaussian state for each gauge configuration, allowing an exact evaluation of the fermionic part of expectation values. We show that this method captures the system’s physics well, even in sign-problem regimes. Using two fermion flavors at nonzero isospin chemical potential as a test case, we identify density-induces phase transitions in the mass–chemical potential plane and demonstrate access to regimes beyond the reach of conventional Lagrangian-based Monte Carlo methods.