Genetic Optimization of Ansatz Expressibility for Enhanced Variational-Quantum-Algorithm Performance
The Research group of Dr. Ankur Raina [Department of Electrical Engineering and Computer Science, IISER Bhopal] developed a genetic algorithm (GA)-based framework for automatically designing quantum circuit ansatzes for variational quantum algorithms (VQAs). VQAs are a leading approach to extracting useful computation from today’s noisy, near-term quantum devices, but their performance critically depends on the design of the parametrized quantum circuit, or ansatz. An ideal ansatz must be expressive enough to represent target quantum states while remaining shallow and trainable. The proposed framework evolves circuit architectures through mutation, selection, and crossover, guided by an expressibility metric based on the Jensen–Shannon divergence from the Haar-random distribution. The evolved ansatzes consistently achieve high expressibility at any target circuit depth. Crucially, a single GA-designed circuit can be reused across multiple quantum chemistry and spin-model problems without repeating the search, reducing classical overhead. Benchmarked against established methods—UCCSD and ADAPT-VQE—the GA ansatz achieves comparable energy accuracy on H?, LiH, BeH?, and H?O using circuit depths up to 30× shallower than UCCSD. Published in Physical Review A (2026). For more details, kindly visit https://doi.org/10.1103/812w-ytnb
