Deterministic Quantum Search via Recursive Oracle Expansion

Abstract: We introduce a novel deterministic quantum search algorithm that provides a practical alternative to conventional probabilistic search approaches. Our scheme eliminates the inherent uncertainty of quantum search without relying on arbitrary phase rotations, a key limitation of other deterministic methods. The algorithm achieves certainty by recursively expanding the base oracle so that it marks all states prefixed by the same two bits as the target, encompassing exactly one-quarter of the search space. This enables a step-by-step reduction of the superposition until the target state can be measured with certainty. The algorithm achieves deterministic success with a query complexity of $O(N^{\log_2(3)/2}) \approx O(N^{0.7925})$, falling between Grover's $O(\sqrt{N})$ scaling and the classical $O(N)$. Our approach relies exclusively on two-qubit nearest-neighbour diffusion operators, avoiding global diffusion entirely. We show that, despite the increased query complexity, this design reduces the total number of two-qubit gates required for diffusion by more than an order of magnitude for search spaces up to at least 18 qubits, with even greater advantages on hardware with limited qubit connectivity. The scheme's inherent determinism, reliance on simple nearest-neighbour, low-depth operations, and scalable recursive structure make it well-suited for hardware implementation. Additionally, we show that the algorithm naturally supports partial database search, enabling deterministic identification of selected target bits without requiring a full search, further broadening its applicability.