Dual-channel multi-product formulas

Product formula (PF) is a promising approach for simulating quantum systems on near-term digital quantum computers. Achieving a desired simulation precision typically requires a polynomially increasing number of Trotter steps, which remains challenging due to the limited capabilities of current quantum hardware. To alleviate this issue, algorithmic error-mitigation techniques such as the multi-product formula (MPF) have been introduced to suppress Trotter errors under restricted hardware resources. In this work, we propose a dual-channel MPF protocol that utilizes the average of a PF and its reversed-sequence counterpart without increasing circuit depth. Our method enables the target simulation precision to be reached with approximately half the circuit depth compared to conventional MPF schemes. This provides substantial error reduction on noisy hardware while implementing a well-conditioned MPF. We also demonstrate our algorithm through numerical simulations under a noise model, showing that our proposal achieves significantly smaller errors.