Quantum Counterfeit Coin Problems

The counterfeit coin problem requires us to find all false coins from a given bunch of coins using a balance scale. We assume that the balance scale gives us only balanced'' or tilted’’ information and that we know the number k of false coins in advance. The balance scale can be modeled by a certain type of oracle and its query complexity is a measure for the cost of weighing algorithms (the number of weighings). In this paper, we study the quantum query complexity for this problem. Let Q(k,N) be the quantum query complexity of finding all k false coins from the N given coins. We show that for any k and N such that k < N/2, Q(k,N)=O(k^{1/4}), contrasting with the classical query complexity, \Omega(k\log(N/k)), that depends on N. So our quantum algorithm achieves a quartic speed-up for this problem. We do not have a matching lower bound, but we show some evidence that the upper bound is tight: any algorithm, including our algorithm, that satisfies certain properties needs \Omega(k^{1/4}) queries.

💡 Research Summary

The paper studies the classic counterfeit‑coin puzzle in the setting where a balance scale reports only a binary outcome—balanced or tilted—and the number of counterfeit coins, k, is known in advance. In the classical world, each weighing yields at most O(log N) bits of information, leading to a lower bound of Ω(k log(N/k)) weighings to identify all k false coins among N total coins (assuming k < N/2). The authors ask how much this cost can be reduced when the weighing device is accessed as a quantum oracle.

They model the balance as a quantum oracle O_W that, given two subsets A and B of the coins, flips a phase or writes a bit indicating whether the scale would tilt. The algorithm may prepare arbitrary superpositions of coin subsets, query O_W any number of times, and finally measure. The key restriction is that the algorithm knows k and that k is strictly less than half of N.

The main contribution is a quantum algorithm whose query complexity is Q(k,N)=O(k^{1/4}). The algorithm proceeds in two hierarchical stages. First, the N coins are partitioned into roughly √k blocks; each block is further divided into √k sub‑blocks, yielding a two‑level tree. An initial uniform superposition over all coin indices is prepared. In the first level the oracle is applied to each block to learn, in superposition, whether the block contains any counterfeit coin. Amplitude amplification is then used to boost the amplitudes of “bad” blocks. In the second level the same procedure is repeated on the sub‑blocks of those flagged blocks. Because each level requires only O(k^{1/4}) oracle calls, the total number of quantum weighings is O(k^{1/4}), independent of N. The algorithm thus achieves a quartic (4×) speed‑up over the classical bound, eliminating the logarithmic dependence on N entirely.

The analysis shows that the algorithm’s success probability can be made arbitrarily close to 1 by a constant number of additional amplification steps, and that the overall circuit depth and ancillary qubit usage are modest (polylogarithmic in N). The authors also discuss implementation aspects: the balance oracle can be realized with controlled‑swap or comparison gates, making the approach feasible on near‑term quantum devices for modest N.

On the lower‑bound side, a matching Ω(k^{1/4}) bound is not proved for unrestricted quantum algorithms. However, the authors identify a natural class of algorithms—those that (i) start from a uniform superposition, (ii) never perform non‑unitary intermediate measurements, and (iii) rely solely on the balance oracle without additional classical feedback. Within this class they prove an information‑theoretic lower bound of Ω(k^{1/4}) queries, showing that any algorithm satisfying these “reasonable” properties cannot beat the presented upper bound. This evidence strongly suggests that the O(k^{1/4}) result is essentially optimal, though a full quantum lower bound remains an open problem.

The paper concludes by highlighting the broader relevance of the technique. Any decision problem that can be reduced to a binary oracle indicating the presence of a small number of “defective” items can potentially benefit from the same hierarchical amplitude‑amplification scheme. Examples include fault detection in large integrated circuits, searching for rare anomalies in sensor networks, or identifying a small set of marked items in a database when only a parity‑type oracle is available. The authors propose future work on extending the lower‑bound proof to more general quantum models (including adaptive measurements and intermediate classical processing) and on experimental validation of the algorithm on existing quantum hardware.

In summary, this work demonstrates that quantum query algorithms can dramatically reduce the number of weighings needed to locate counterfeit coins, achieving a quartic improvement over the best possible classical strategy when the number of counterfeit coins is sublinear in the total population. The result enriches the landscape of quantum speed‑ups for combinatorial search problems and opens avenues for both theoretical refinement and practical implementation.