Integrating Self-initialized Local Thermalizing Lindblad Operators for Variational Quantum Algorithm with Quantum Jump: Implementation and Performance

Quantum computing holds great potential for simulating quantum systems, such as molecular systems, due to its inherent ability to represent and manipulate quantum states. However, simulating nonunitary dynamics of open quantum systems on a quantum computer is still challenging. Here, we present a scheme denoted as VQS-QJ-LTLME, which adopts the trajectory average of quantum jump Monte Carlo wave function variational evolution to evolve the system's density matrix and local thermalizing Lindblad operators to describe the system-environment interactions. This combination allows the quantum circuit to be initialized after a period of evolution, thereby eliminating the accumulation of errors. The VQS-QJ-LTLME algorithm requires only log_2 qubits for system size $n$ and holds a time complexity of O(Tn^3), making it particularly suitable for operation on noisy intermediate-scale quantum devices. To accelerate the Monte Carlo sampling process in VQS-QJ-LTLME, we introduce an efficient sampling method, named No-evolution sampling (NES). The VQS-QJ-LTLME aided by NES requires simulating only (n^3+1) trajectories on quantum computers, and then makes 10^6 ~ 10^7 samplings on classical computers in a few seconds. To demonstrate the performance of the VQS-QJ-LTLME algorithm, we simulate the dynamics of a two-level spin-boson model and a four-level reduced Fenna-Matthews-Olson system with real superconducting quantum computers and classical simulators. It is shown that the simulations produced by VQS-QJ-LTLME algorithm align closely with those obtained from the purely classical methods.