Author
Fangqi Lou
Recent research
- AI & ComputingOpen access
A Lovász Theta Bound for a Graph Saturation Parameter
For a finite simple graph G on n vertices, let μ(G) be the minimum trace of a real symmetric matrix X satisfying 0 ≼ X ≼ I, Xᵢⱼ = 0 on the edges of G, and ⟨J,X⟩ = n. We prove that μ(G) ≤ ϑ(complement G) for every G, answering affirmatively question (iii) in Section 7 of Carli Sil...
- AI & ComputingOpen access
Shadow-flow rigidity of D_n polytopes and quartic descent at the 24-cell
For every n >= 4, we prove that the D_n root polytope and its polar have only globally linear odd admissible vertex speeds in every nonzero direction. Admissibility means affine compatibility on every facet transverse to that direction. For the regular 24-cell, every fixed nontri...
- AI & ComputingOpen access
Shadow-flow rigidity of D_n polytopes and quartic descent at the 24-cell
For every n >= 4, we prove that the D_n root polytope and its polar have only globally linear odd admissible vertex speeds in every nonzero direction. Admissibility means affine compatibility on every facet transverse to that direction. For the regular 24-cell, every fixed nontri...
- AI & ComputingOpen access
Double-star labellings of defect one can require arbitrarily many edge-difference changes
For a bijective labelling of an n-vertex tree by 0,...,n-1, define its defect as n-1 minus the number of distinct absolute edge differences. We compare two labellings by counting the fixed edges on which their differences disagree. For every integer s >= 2, we construct a double...
- AI & ComputingOpen access
A Lovász Theta Bound for a Graph Saturation Parameter
For a finite simple graph G on n vertices, let μ(G) be the minimum trace of a real symmetric matrix X satisfying 0 ≼ X ≼ I, Xᵢⱼ = 0 on the edges of G, and ⟨J,X⟩ = n. We prove that μ(G) ≤ ϑ(complement G) for every G, answering affirmatively question (iii) in Section 7 of Carli Sil...
- AI & ComputingOpen access
Bruhat Monotonicity of the Pan–Skandera–Wang Bijection
Pan, Skandera, and Wang defined a bijection from permutations preserving two balanced consecutive blocks to permutations preserving the parity of every position, and conjectured that each input lies below its image in the strong Bruhat order. We prove Conjecture 7.8 of arXiv:2606...
- AI & ComputingOpen access
Double-star labellings of defect one can require arbitrarily many edge-difference changes
For a bijective labelling of an n-vertex tree by 0,...,n-1, define its defect as n-1 minus the number of distinct absolute edge differences. We compare two labellings by counting the fixed edges on which their differences disagree. For every integer s >= 2, we construct a double...
- AI & ComputingOpen access
Bruhat Monotonicity of the Pan–Skandera–Wang Bijection
Pan, Skandera, and Wang defined a bijection from permutations preserving two balanced consecutive blocks to permutations preserving the parity of every position, and conjectured that each input lies below its image in the strong Bruhat order. We prove Conjecture 7.8 of arXiv:2606...
- Engineering & TechnologyOpen access
The Mean-Field Limit of Online Rademacher Vector Balancing
Let X₁, …, Xₙ be independent uniform vectors in {−1, 1}ⁿ. At time k, a sign is chosen after observing X₁, …, Xₖ. We identify the limit of the minimum expected terminal maximum norm of n^(−1/2) ∑ₖ₌₁ⁿ εₖXₖ with the mean-field control value introduced by Fiedler, Jackson, Lacker, an...
- AI & ComputingOpen access
Positive Lower Natural Density Does Not Ensure Carleson Frame Sampling
We construct a locally finite, strictly increasing real sampling sequence with lower natural density 1/3 and upper natural density 4/3 for which a sampled Carleson orbit is Bessel but has no positive frame lower bound. The diagonal spectrum is fixed explicitly as z_j = exp(-2^(-j...
- Engineering & TechnologyOpen access
The Mean-Field Limit of Online Rademacher Vector Balancing
Let X₁, …, Xₙ be independent uniform vectors in {−1, 1}ⁿ. At time k, a sign is chosen after observing X₁, …, Xₖ. We identify the limit of the minimum expected terminal maximum norm of n^(−1/2) ∑ₖ₌₁ⁿ εₖXₖ with the mean-field control value introduced by Fiedler, Jackson, Lacker, an...
- AI & ComputingOpen access
Positive Lower Natural Density Does Not Ensure Carleson Frame Sampling
We construct a locally finite, strictly increasing real sampling sequence with lower natural density 1/3 and upper natural density 4/3 for which a sampled Carleson orbit is Bessel but has no positive frame lower bound. The diagonal spectrum is fixed explicitly as z_j = exp(-2^(-j...