Uniform mixing and $ε$-uniform mixing on cycles

Abstract: We study continuous-time quantum walks on cycles. We prove two complementary results. Firstly, the cycle $C_9$ does not admit uniform mixing at any time. Using the similar idea and Dickson polynomials, we prove that $C_{15}$ does not admit uniform mixing at any time neither. Secondly, for every prime $p$, we show that the cycle $C_{p<sup>2}$ admits $ε$-uniform mixing.

• The paper presents that uniform mixing fails for cycles C9 and C15, confirmed through algebraic decomposition and Gröbner basis computations.
• It extends the analysis to ε-uniform mixing by proving that cycles of prime-square order (C(p²)) can arbitrarily approach a uniform distribution.
• The findings impact quantum algorithms and spectral graph theory by establishing strong algebraic criteria for perfect quantum mixing in cycles.

Uniform Mixing and $ε$-Uniform Mixing on Cycles: Main Contributions

The paper "Uniform mixing and $ε$-uniform mixing on cycles" (2607.04207) delivers a rigorous analysis of mixing properties of continuous-time quantum walks (CTQWs) on cycle graphs, specifically $C_n$. The central results are as follows:

• It is formally established that $C_9$ and $C_{15}$ do not exhibit uniform mixing, closing previously open cases after prior partial results for other values of $n$.
• The paper extends the theory of $ε$-uniform mixing, proving that for any prime $p$, the cycle $C_{p^2}$ always admits $ε$-uniform mixing, i.e., the quantum walk can get arbitrarily close to uniform distribution in terms of the Frobenius norm at some time $ε$0.

These results consolidate the understanding around Ahmadi et al.'s conjecture on the rarity of uniform mixing for cycles, as well as advancing the classification of cycles that allow approximate mixing.

Uniform Mixing Criteria and Algebraic Characterization

Uniform mixing on a graph $ε$1 at time $ε$2 means the transition matrix $ε$3 (where $ε$4 is the adjacency matrix) is flat up to normalization; each entry has squared modulus $ε$5 for $ε$6 vertices. The authors provide a decomposition of this condition using the circulant structure of the adjacency matrix of $ε$7. In particular, they show that uniform mixing occurs iff

$ε$8

where $ε$9 and $C_n$0 are eigenvalues.

To facilitate calculations, the eigenstructure is recast using Dickson polynomials, allowing algebraic manipulation of the relations between characteristic roots for different $C_n$1. Lemmas concerning Vandermonde matrices and linear independence underlie these reductions.

Non-Existence of Uniform Mixing on $C_n$2 and $C_n$3

The paper precisely settles the two significant open instances:

• $C_n$4: Using the decomposed algebraic conditions above, the analysis reduces to a polynomial system in variables derived from the eigenvalues' exponentials. By implementing Gröbner basis computation, it is shown the only solutions violate unimodularity necessary for quantum walks, implying uniform mixing never occurs for $C_n$5.
• $C_n$6: By employing the minimal polynomial for $C_n$7 and expanding explicit expressions for all $C_n$8 in terms of a small generating set, the authors again derive a polynomial system whose only solutions are impossible for unimodulus values. Computation with Gröbner bases confirms the non-occurrence of uniform mixing.

Both proofs are computationally explicit, leveraging algebraic properties unique to these cases and confirming, with no reliance on approximate arguments, that $C_n$9 and $C_9$0 fail to admit uniform mixing at any real time.

$C_9$1-Uniform Mixing on Cycles of Prime-Square Order

Addressing cycles that may not admit exact uniform mixing, the paper formally defines $C_9$2-uniform mixing: for any $C_9$3, it is possible to find $C_9$4 such that $C_9$5 (Schur product, entrywise) is within $C_9$6 of the perfectly flat all-ones matrix in Frobenius norm.

• The main positive result: If $C_9$7 is prime, $C_9$8 admits $C_9$9-uniform mixing.
• The proof utilizes explicit construction of unitary, flat matrices $C_{15}$0 approximating the time-evolved quantum walk, with eigenvalues approximated by controlled roots of unity.
• The main technical ingredient is an application of Kronecker's theorem, ensuring that the linear independence of appropriately normalized eigenvalue sets implies density of the walk's evolution in the relevant torus, and hence the ability to approach any phase configuration (and thus the flat matrix) arbitrarily closely.

This advances earlier work, which only guaranteed $C_{15}$1-uniform mixing for $C_{15}$2, to include all prime-squared cycles.

Implications and Directions for Future Research

The results have theoretical and practical implications:

• Quantum Algorithms: The negative results strengthen the understanding that uniform (perfect) mixing in CTQWs—necessary for certain quantum algorithms and perfect sampling schemes—is extremely limited on cycles, essentially occurring only for trivial small cases ($C_{15}$3, $C_{15}$4).
• Spectrum-Driven Approaches: The positive result for $C_{15}$5-uniform mixing on $C_{15}$6 opens up new directions for leveraging spectral properties and algebraic independence in CTQW mixing, suggesting explicit design criteria for graphs with near-uniform quantum mixing.
• Computational Algebraic Techniques: The device of using Gröbner bases for certifying impossibility results in quantum walk theory may have broader applications, especially for families of graphs with high symmetry.

As open problems, the authors conjecture (in keeping with a proposal by Mullin) that if any graph admits uniform mixing at $C_{15}$7, then $C_{15}$8 must be a root of unity—hinting at a strong link between algebraic properties of the graph spectrum and uniform mixing. Additionally, they point out the challenge in extending the $C_{15}$9-uniform mixing construction to cycles of order $n$0 for $n$1.

Conclusion

The paper provides a mathematically rigorous account of both uniform and approximate ($n$2-) uniform mixing in continuous-time quantum walks on cycles. It fully settles two previously unsettled cycle lengths, gives explicit algebraic conditions for uniform mixing, and expands the class of cycles known to admit $n$3-uniform mixing. Theoretical consequences tie the occurrence of mixing closely to algebraic and arithmetic properties of the cycle order, and suggest new questions in both graph theory and quantum computation.