Toward High-Throughput Virtual Screening via Quasi-Adiabatic Quantum Geometry Optimization Workflow
This work addresses the challenge of molecular geometry optimization at post-Hartree–Fock levels of theory using quantum computing methodologies. We present a novel potentially quasi-adiabatic approach based on the Variational Quantum Eigensolver (VQE) algorithm for efficient exploration of potential energy surfaces. The method combines quantum-circuit emulators with classical optimization routines and uses the Hellmann–Feynman theorem, including Pulay corrections, together with a finite-difference scheme to compute forces. Our implementation uses GPU-accelerated quantum-circuit operations within a parallelized framework, achieving computational efficiency by warm-starting the wavefunction and reoptimizing it at each subsequent geometry-optimization step. Benchmark results on small molecular systems demonstrate the feasibility of this hybrid quantum-classical approach, with detailed comparisons against established quantum chemistry methods, including CASCI and DFT optimizations. We discuss the current limitations, computational advantages, and potential future developments of this methodology in the context of quantum chemistry for drug discovery applications.