Constants defined by an optimisation problem — the best bound in an inequality, the extremal value of a construction — whose exact value is unknown. Terence Tao and contributors keep the best known lower and upper bounds. An explicit construction that beats a bound is checked by recomputing it; new records should also be reported upstream.
Source: Crowdsourced repository of optimization constants (Terence Tao and contributors). Licence: Apache License 2.0.
Let x_1,…,x_10 be very general points of ℙ^2, and let π:X→ ℙ^2 be the blow-up of ℙ^2 at these points. Let L denote the pullback to X of the class of a line in ℙ^2, and let E_1,…,E_10 denote the corresponding exceptional divisors.
In harmonic analysis, for λ > 0 let T^λ denote the Bochner–Riesz operator on ℝ^3, initially defined for Schwartz functions f ∈ S(ℝ^3) by T^λ f(x) := ∫_ℝ^3 (1-lvert ξ rvert^2)_+^λ widehatf(ξ)e^ix· ξ dξ, where widehatf denotes the Fourier transform of f and (t)_+ := max\t,0\.
C_3c = SD(\0,1,2,∞\;-1) is the least exponent such that one has the inequality |A stackrelG- B| ≤ max(|A|, |B|, |A stackrelG+ B|, |A stackrelG+ 2B|)^C_3c whenever A, B are finite subsets of reals and G ⊂ A × B, where A stackrelG± rB := a ± rb: a ∈ A, b ∈ B.
C_5a is the smallest constant such that Sidon sets in \1,…,N\ have cardinality N^1/2 + (C_5a + o(1))N^1/4.
The ambidextrous moving sofa constant C_41b asks for the maximum area of a sofa, as defined in C_41a that can navigate both left and right corners inside a Z-shaped corridor of width 1, where the corners are sufficiently far apart.
C_1a is the largest constant for which one has max_-1/2 ≤ t ≤ 1/2 ∫_ℝ f(t-x) f(x) dx ≥ C_1a (∫_-1/4^1/4 f(x) dx)^2 for all non-negative f : ℝ → ℝ.
Quantum Max Cut is the quantum analog of Max Cut. Given a graph G = (E,V), it asks for the maximum eigenvalue of H_G = Σ_(ij) ∈ E (I - X_iX_j - Y_iY_j - Z_i Z_j), where X_i, Y_i, Z_i are the Pauli matrices acting on the i'th tensor factor and trivially on all other coordinates.
For integers n ≥ 2 and t ≥ 1, an n × t partial Hadamard matrix is a matrix with entries in \± 1\ whose rows are pairwise orthogonal. Let N_n,t denote the number of such matrices. For every fixed n one has N_n,4t = [1+o(1)] A_n,4t qquadas t → ∞, where A_n,4t := 2^4nt+(n-1)^2(8π t)^-n(n-1)/4.
Let α be a nonzero algebraic number of degree d, with minimal polynomial over ℤ f(X)=a_dΠ_i=1^d (X-α_i), where a_d>0 and α_1,…,α_d are the conjugates of α. Define the Mahler measure of α by M(α) := a_dΠ_i=1^d max1,lvert α_irvert.
For each integer n ≥ 1, let S_n be the symmetric group on n letters. Over a base field k, the essential dimension ed_k(S_n) is the smallest integer d such that the general degree-n polynomial x^n + a_1 x^n-1 + ⋯ + a_n can be reduced to a d-parameter form by a Tschirnhaus transformation.
For each integer d ≥ 3, let ℓ_0(d) denote the maximal number of lines contained in a smooth surface of degree d in ℙ^3_ℂ. We define C_76 := limsup_d→∞ℓ_0(d)/d^2. The constant C_76 measures the quadratic growth rate of the maximal line count on smooth complex degree-d surfaces.
In harmonic analysis, the Beurling–Ahlfors transform B (also called the Ahlfors–Beurling operator) is the singular integral operator on L^p(ℂ), 1<p<∞, defined by Bf(z) = -1/π p.v.∫_ℂ f(w)/(z-w)^2 dm(w) = -1/π lim_ε→ 0^+∫_lvert w-zrvert>varepsilonf(w)/(z-w)^2 dm(w), where dm is Lebesgue measure on…
Let D=\z∈ℂ:lvert zrvert<1\. Following standard notation, let F be the class of holomorphic functions f:D→ℂ normalized by lvert f'(0)rvert=1 (equivalently, after rotation, f'(0)=1).
Degree at most d functions f:lbrace ± 1rbrace^n→ℝ have Fourier–Walsh expansion f(x)=Σ_S⊆ [n], |S|≤ d widehat f(S) x^S, x^S:=Π_i∈ Sx_i, [n]:=lbrace 1,…,nrbrace. For d∈ℕ set p_d:=2d/d+1.
Let D^d := z=(z_1,…,z_d)∈ℂ^d: lvert z_1rvert,…,lvert z_drvert<1 be the unit polydisc, and let the Schur class S_d be the set of analytic functions f:D^dtoD.
Let p_n denote the n-th prime. The bounded prime gap constant is C_88a = H_1 := liminf_n → ∞ (p_n+1 - p_n), the least limit point of the sequence of gaps between consecutive primes.
Let Ω⊂ℂ be simply connected with at least two boundary points in the extended complex plane, and let φ:ΩtoD be a conformal map. Brennan's conjecture states that ∫_Ωlvert φ'(z)rvert^p dx dy < ∞ qquadwhenever 4/3<p<4.
C_16 = L is the smallest constant for which the sharp Brezis–Gallouet inequality ‖u‖_L^∞(T^2)^2 ≤ 1/4π ‖∇ u‖_L^2(T^2)^2 Bigl[lnδ(u) + lnbigl(1+lnδ(u)bigr) + LBigr] holds for all zero-mean functions u ∈ H^2(T^2) with sufficiently large frequency ratio δ(u) := ‖Δ u‖_L^2(T^2)^2/‖∇ u‖_L^2(T^2)^2.
C_81a, Brun's Constant, is the sum of the reciprocals of the twin primes.
A Dirichlet character of level q is an arithmetic function χ that is multiplicative, is defined by a character on (ℤ/qℤ)^ast on integers coprime to q, and is 0 on integers not coprime to q.
In ℝ^d (d≥ 1), let M_d denote the centered Hardy–Littlewood maximal operator associated to cubes, defined by M_d f(x) := sup_r>0 1/lvert Q(x,r)rvert∫_Q(x,r) lvert f(y)rvert dy, where Q(x,r) is a closed ℓ_∞ ball of radius r and center x in ℝ^d, that is, a closed cube centered at x, with sides…
Let λ_n,2 be the random variable assigning two uniformly random binary strings of length n the length of their longest common subsequence. Then C_31a is the (well-defined) limit C_31a := lim_n → inftyE[λ_n,2]/n.
C_8 = R is the least constant such that there are no zeroes σ+it of the Riemann zeta function with lvert t rvert ≥ 2 and σ > 1 - 1/R log lvert t rvert.
The constant term of one-shot channel simulation [HJMR07], [BG14], [LEG18], [Li25] is given as (we use the definition in [Li25]) C_32=limsup_t→∞(sup_p_X,Y: I(X;Y)=t inf_p_S|X,Y: I(X;S)=0H(Y|S)-t-log_2t), where H(Ylvert S)=H(Y,S)-H(S) is the conditional entropy (in bits), and I(X;Y)=H(X)+H(Y)-H(X,Y)…
In topological graph theory, a thrackle is a drawing of a finite graph in the plane in which every pair of edges meets precisely once, either at a common endpoint or at a proper crossing.
In zero-sum theory, the Davenport constant D(G) of a finite abelian group G is defined as the smallest integer l∈ℕ such that every sequence S over G of length lvert Srvert≥ l has a non-empty zero-sum subsequence.
Let d(n) be the divisor function. The Dirichlet divisor problem concerns the error term Δ(x) := Σ_n≤ x d(n) - x(log x + 2γ -1), where γ is Euler's constant. <a href="#Tsa2010-def-Delta">[Tsa2010-def-Delta]</a> Define the divisor-problem exponent α := infBigla≥ 0: Δ(x)=O(x^a+ε) for all ε>0Bigr.
In algebraic complexity theory, for each real k ≥ 0, let ω(k) denote the exponent for multiplying an n × n^k matrix by an n^k × n matrix. We define α := supk ≥ 0 : ω(k) = 2.
Let Λ denote the von Mangoldt function. For coprime positive integers a,q, define ψ(x;q,a) := Σ_n≤ x, n≡ a (mod q) Λ(n).
For any transcendental entire function f(z)=Σ_n≥ 0 a_n z^n, define . M(r,f):=max_|z|=r|f(z)|, μ(r,f):=max_n≥ 0|a_n| r^n. Following [Er1961], define β(f):=liminf_r→∞μ(r,f)/M(r,f). We define C_51 = B to be the supremum of β(f) over all transcendental entire functions f.
A family of three distinct sets A,B,C is a 3-sunflower (or Δ-system) if A∩ B = A∩ C = B∩ C. A family of sets is sunflower-free if it contains no 3-sunflower (equivalently, no sunflower of any size ≥ 3). Let [n]:=\1,2,…,n\ and let f(n) denote the maximum size of a sunflower-free family F⊆ 2^[n].
Let overlineℚ be the set of all algebraic numbers. The naïve height h : overlineℚ → ℝ is defined as follows. Let α ∈ overlineℚ and let P(x) be an irreducible primitive polynomial with integers coefficients such that P(α)=0. Let n be the degree and a be the leading coefficient of P(x).
Let p_n denote the n-th prime and, for m ≥ 1, write H_m := liminf_n → ∞ (p_n+m - p_n) for the least limit point of the gaps between primes m apart.
Let H be an infinite-dimensional complex Hilbert space and let B(H) be the Banach algebra of bounded operators on H, equipped with the operator norm.
C_17 is the limit (if it exists) of R(k)^1/k as k → ∞, where the diagonal Ramsey number R(k) is the smallest integer n such that every red/blue colouring of the edges of the complete graph K_n contains a monochromatic copy of K_k.
The Falconer distance problem threshold C_34 = s_Δ(ℝ^2) in the plane is defined as s_Δ(ℝ^2) : :=\ infBigl s∈[0,2] : ∀ compact E⊂ℝ^2,\ dim_H(E)>s Longrightarrow lvertΔ(E)rvert>0 Bigr.
Let E⊂ ℝ^2 be a planar set. The Favard length of E is defined by Fav(E) := 1/π∫_0^π lvert Proj R_θ Ervert dθ, where Proj is orthogonal projection to the horizontal axis and R_θ is rotation by angle θ.
A convex body K⊂ℝ^d is hollow (lattice-free) with respect to a lattice Λ if int(K)∩Λ=∅. <a href="#CS2019-hollow-def">[CS2019-hollow-def]</a> For a hollow body, the lattice width is w(K) := min_u∈ℤ^d∖0 (max_x∈ Ku· x-min_x∈ Ku· x).
Let f:\-1,1\^n→\-1,1\ be a Boolean function with Fourier expansion f(x)=Σ_S⊆[n]hat f(S)χ_S(x). Its spectral entropy is H(hat f^2) := Σ_S⊆[n]hat f(S)^2log_21/hat f(S)^2, and its total influence is Inf(f) := Σ_S⊆[n]hat f(S)^2 lvert Srvert.
Let r(N) be the maximum size of a subset A⊂\1,…,N\ with no non-zero square differences a-b=n^2.Then C_4b is the least constant such that r(N) ≤ N^C_4b+o(1).
Let N(t) := \#(m,n)∈ℤ^2: m^2+n^2≤ t^2 be the number of integer lattice points inside the (closed) disk of radius t centered at the origin. The Gauss circle problem is to find the smallest exponent θ such that, for every ε>0, N(t) = π t^2 + O(t^θ+ε).
C_43 is defined as the infimum of the ratio of the length of the Steiner Minimal Tree to the length of the Euclidean Minimum Spanning Tree over all finite sets of points V ⊆ ℝ^2: C_43 = inf_VL_S(V)/L_M(V), where L_S(V) and L_M(V) denote the lengths of Steiner Minimal Tree and Minimum Spanning Tree…
We define C_56 = δ_2 to be the smallest real number δ ≥ 0 such that the following uniform bound toward the Generalized Ramanujan Conjecture holds.
Let f be a convex function with 1-Lipschitz gradient. We assume black-box access to the function and its gradient. Gradient descent will converge to a global minimum with an appropriate choice of _step size_ s: x_k+1 := x_k - s· ∇ f(x_k). In general, s can be chosen to vary with the step k.
C_39=H_3 is the Hadwiger covering number in dimension 3, which can also be formulated in terms of illumination of the boundary.
C_33=A(2) is the Ihara constant over 𝔽_2. <a href="#DM2013-def-Aq">[DM2013-def-Aq]</a> For each integer g≥ 1, let N_2(g) := maxbigl\#X(𝔽_2) : X/𝔽_2 a smooth projective geometrically integral curve of genus gbigr. <a href="#DM2013-def-Nqg">[DM2013-def-Nqg]</a> Then A(2) := limsup_g→inftyN_2(g)/g.
Let G = (g_ij) be an M × N random matrix with independent standard Gaussian entries, and let Z(G) := | σ ∈ -1,1^N : G σ ≥ 0 coordinatewise |. This is the zero-margin binary (or Ising) perceptron. Write M = ⌊ α N ⌋. Define C_80 to be the infimum of all α > 0 such that ℙ(Z(G) > 0) → 0 as N → ∞.
C_3b = SD(\0,1,∞\;-1) is the least exponent such that one has the inequality |A stackrelG- B| ≤ max(|A|, |B|, |A stackrelG+ B|)^C_3b whenever A, B are finite subsets of reals and G ⊂ A × B, where A stackrelG± B := a ± b: a ∈ A, b ∈ B.
C_24 is the Komlós discrepancy constant (often denoted K). For a real matrix A∈ℝ^m× n, define its (sign) discrepancy by disc(A) := min_x∈-1,1^n ‖Ax‖_∞. For each n≥ 1, define the dimension-n Komlós discrepancy K_n := supdisc(A): A∈ℝ^n× n and ‖A_ast j‖_2≤ 1 for all columns j.
Let D:=\z∈ℂ:lvert zrvert<1\. The Bergman space A^2(D) consists of analytic functions f on D with lVert frVert_2 := (1/π∫_D lvert f(z)rvert^2 dA(z))^1/2 < ∞, where dA(z) denotes the Lebesgue area measure. For c∈(0,1), write A(c,1) := z∈ℂ: c<lvert zrvert<1.
Let D=\z∈ℂ:lvert zrvert<1\ and let F be the class of holomorphic functions f:D→ℂ normalized by f'(0)=1. <a href="#BS2023-def-F">[BS2023-def-F]</a> For finF, let L_f denote the radius of the largest disk contained in f(D).
C_13b = a is the infimal area of a convex planar set Ω that can cover a congruent copy of every convex planar set of diameter 1.
Let f(x)=Σ_i=0^n a_i x^i = a_nΠ_i=1^n (x-α_i) be a polynomial with complex coefficients. The Mahler measure of f is M(f) := |a_n|Π_i=1^n max1,|α_i|.
Define the infimal exponent μ_ζ by μ_ζ := infBiglθ≥ 0: lvertζ(1/2+it)rvert≪_ε(1+lvert trvert)^θ+ε for all ε>0Bigr. We define C_62a := μ_ζ, the Lindelof (pointwise growth) exponent for ζ(1/2+it).
For integers q≥ 2 and a with gcd(a,q)=1, let P(a,q) denote the least prime in the arithmetic progression a bmod q. <a href="#Xyl2011-def-Paq">[Xyl2011-def-Paq]</a> Linnik's theorem asserts that there exist constants C,L>0 such that P(a,q) ≤ C q^L (gcd(a,q)=1), uniformly for all q≥ 2.
Let K⊂ℝ^n be a centrally symmetric convex body (compact, convex, with non-empty interior) satisfying K=-K. Its polar body is K^∘ := y∈ℝ^n: ⟨ x,y⟩ ≤ 1 for all x∈ K. The volume product of K is vp(K) := Vol_n(K) Vol_n(K^∘).
For a number field K, let Δ_K denote the absolute value of its discriminant and let [K:ℚ] denote its degree. The root discriminant of K is rd(K) := Δ_K^1/[K:ℚ].
C_18 is the least constant such that, whenever A is a subset of 𝔽_2^n with lvert A+Arvert ≤ Klvert Arvert, then A can be covered by K^C_18+o(1) cosets of a subspace of cardinality at most lvert Arvert, where the limit o(1) is with respect to the limit K → ∞.
Let f:0,1^n→0,1 be a Boolean function. Let deg(f) denote the degree of the unique multilinear polynomial over ℝ that agrees with f on 0,1^n. A variable x_i is relevant if f depends on it (equivalently: x_i appears in some monomial with nonzero coefficient in the multilinear representation of f).
C_27b is the highest possible chromatic number for any biplanar graph.
In the symmetric metric traveling salesman problem, one is given a complete graph K_n=(V,E) with a nonnegative symmetric cost function c:E→ ℝ_≥ 0 satisfying the triangle inequality. For S⊆ V, let δ(S) denote the set of edges with exactly one endpoint in S, and write δ(v):=δ(\v\).
C_13a is the infimal area of a convex domain Ω that can contain a rigid motion (translation + rotation; no reflections) of every planar arc (curve, or "worm") of length 1.
The moving sofa constant C_41a=A is the maximum area of a connected, rigid planar shape that can maneuver through an L-shaped corridor of unit width. The corridor is formed by two semi-infinite strips of width 1 meeting at a right angle.
For integers m,n≥ 1, let B_ℝ,m(n) be the smallest constant such that every m-linear form T:(ℓ_∞^n)^m → ℝ satisfies the (multilinear) Bohnenblust–Hille inequality (Σ_j_1,…,j_m=1^n bigl|T(e_j_1,…,e_j_m)bigr|^2m/m+1)^m+1/2m ≤ B_ℝ,m(n) ‖T‖, where ‖T‖:=sup_‖x^(1)‖_∞,…,‖x^(m)‖_∞ ≤…
Let X be an integrable real random variable. We say that X is 1-sub-Gaussian in the tail sense if E[X]=0 quadand ℙ(lvert Xrvert>t)≤ 2e^-t^2/2quadfor all t≥ 0.
Let χ be a primitive Dirichlet character modulo q, and define S(χ) := max_N≤ q lvertΣ_1≤ n≤ Nχ(n)rvert. The Polya-Vinogradov inequality states that S(χ) ≤ c √(q) log q for some absolute constant c. <a href="#BK2020-def-PV">[BK2020-def-PV]</a> For squarefree moduli, define C_72^even (resp.
Given a real matrix A, let its condition number be κ(A):=σ_max(A)/σ_min(A), where σ_min(A) and σ_max(A) denote the smallest and largest singular values of A, respectively (with κ(A)=∞ if σ_min(A)=0).
C_46 is the infimal exponent p such that one has the global bound ‖widehatf dσ‖_L^p(ℝ^3) lesssim_p ‖f‖_L^∞(S^2) qquadfor all f∈ L^∞(S^2).
For subsets K,L⊂ℝ^n, their Minkowski sum is K+L := x+y: x∈ K, y∈ L. In general, one cannot expect a reverse Brunn-Minkowski inequality for arbitrary compact sets, even with a fixed multiplicative constant.
C_45 is the asymptotic density (if it exists) of the set of odd integers that can be expressed as the sum of a prime number and a power of two.
An algebraic integer α of degree d, with conjugates α_1,…,α_d, is totally positive if all of its conjugates are real and strictly positive. Its absolute trace (or trace-to-degree ratio) is overlinetr(α) := tr(α)/deg(α) = 1/dΣ_i=1^d α_i . Let A denote the set of totally positive algebraic integers.
Let Γ⊂ SL_2(ℤ) be a congruence subgroup. Denote by 0=λ_0<λ_1(Γ)≤ λ_2(Γ)≤ ⋯ the eigenvalues of the (non-Euclidean) Laplacian acting on L^2(ΓbackslashH).
Let f:ℂ→ℂ be a polynomial of degree n≥ 2 whose zeroes all lie in the closed unit disk D(0,1)=\z:lvert zrvert≤ 1\. Sendov's conjecture states that if λ_0 is one of these zeroes, then f' has at least one zero in D(λ_0,1). every zero λ_0 of f has a critical point in D(λ_0,1).
Let C_7 denote the cycle graph on 7 vertices. We define C_9 to be the Shannon capacity of mathcal C_7: C_9 := Θ(mathcal C_7), where for a graph G, the Shannon capacity Θ(G) is defined by Θ(G) := sup_n ≥ 1 α(G^boxtimes n)^1/n.
C_5b is the largest constant such that every (4,5)-set of size n (i.e., a set of reals such that every four-element subset determines at least five distinct differences) contains a Sidon set of cardinality C_5bn.
For a finite nonempty subset A of an abelian group, write σ(A) := |A+A|/|A|, δ(A) := |A-A|/|A| for the doubling and difference constants. Ruzsa [Ru96] proved δ ≤ σ^2, and the Plünnecke–Ruzsa inequalities give the converse σ ≤ δ^2 [Bl26].
For a bounded set X⊂ ℝ^n, its diameter is diam(X) := sup‖x-y‖_2: x,y∈ X. Let b(X) be the smallest integer m such that X can be written as a union X = X_1 ∪ ⋯ ∪ X_m with diam(X_i) < diam(X) for all i=1,…,m.
C_14 is the smallest n, such that the value of the busy beaver number BB(n) is undecidable in ZFC (or equivalently ZF). Explicitly, it is the smallest n such that there is a Turing machine with n states for which it cannot be proven in ZFC (assuming ZFC is consistent) whether it halts or not.
C_10c is the least constant K for which one has disc(A) ≤ K√(n) qquadfor all n and all A∈[-1,1]^n× n. Here the discrepancy disc(A) is defined as disc(A) := min_x∈± 1^n ‖Ax‖_∞.
C_36=Δ_4 is the (optimal) sphere packing density in ℝ^4, i.e. the largest fraction of ℝ^4 that can be covered by congruent balls with disjoint interiors.
Let ℤ^2 denote the square lattice graph with vertex set ℤ^2 and edges between nearest neighbors (Euclidean distance 1). A self-avoiding walk (SAW) on a graph G=(V,E) is a walk that visits no vertex more than once.
Let Av_n(1324) be the set of permutations of \1,2,…,n\ that avoid the permutation pattern 1324, and let S_n(1324) := |Av_n(1324)|. <a href="#CJS12-def-Sn">[CJS12-def-Sn]</a> The Stanley–Wilf limit (growth constant) for the pattern 1324 is C_30 := lim_n→∞ bigl(S_n(1324)bigr)^1/n.
For a finite set A ⊂ ℝ write A+A = \ a+b : a,b ∈ A \, AA = \ ab : a,b ∈ A \ for the sumset and product set. The (real) sum-product exponent is C_84b := liminf_n → ∞ min_substackA ⊂ ℝ \ lvert Arvert = n log max(lvert A+Arvert, lvert AArvert)/log n.
C_12 = β_2 is the constant such that the length L_n of the shortest tour through n independent uniform random points satisfies L_n/√(n)→ β_2 almost surely.
Let X_1,X_2,… be i.i.d. real random variables with E X_1 = 0, Var(X_1)=1, and finite third absolute moment β_3 := E|X_1|^3 < ∞. Let S_n := X_1+⋯+X_n/sqrt n, F_n(x):=ℙ(S_n≤ x), and let Φ denote the standard normal distribution function.
Let G be a simple graph. The acyclic chromatic index χ_a'(G) of G is defined to be the least number of colors needed to color the edges of G so that no two edges coincident on the same vertex are homochromatic and there is no cycle whose edges are colored with only two colors.
The complex Grothendieck constant (often denoted K_G^ℂ) is the smallest number C_10b such that, for every m,n≥ 1 and every complex matrix A=(a_ij)∈ℂ^m× n, max_substacku_1,…,u_m∈ S^∞\ v_1,…,v_n∈ S^∞ |Σ_i=1^mΣ_j=1^n a_ij⟨ u_i, v_j⟩| ≤ C_10b\ max_substack|s_1|=⋯=|s_m|=1\ |t_1|=⋯=|t_n|=1…
Let m,n be positive integers and let V be a set of n Boolean variables. By a random formula of density r = m/n, we mean a collection of m clauses selected u.a.r. with replacement from the set of 8C(n, 3) clauses on three distinct variables from V.
Let Q_n = -1,1^n be the Hamming cube (two vertices are adjacent if they differ in exactly one coordinate). For a set A ⊂ Q_n define the function h_A:Q_n→ 0,1,...,n by - h_A(x)=0 if x∉ A; - if x∈ A, then h_A(x) is the number of neighbors of x that lie in the complement A^c.
C_2 is the Crouzeix constant (sometimes denoted Q). It is the smallest constant C such that for every n ≥ 1, every complex matrix A ∈ ℂ^n × n, and every complex polynomial p one has ‖p(A)‖ ≤ C max_z ∈ W(A) |p(z)|, where ‖·‖ is the operator norm induced by the Euclidean norm (i.e.
Let f be a Boolean function on n bits, i.e. f:0,1^n → 0,1 with n≥ 2. For x∈ 0,1^n and 1≤ i≤ n, let x^(i) be x with the i-th bit flipped. The (pointwise) sensitivity of f at x is s(f)(x):=Σ_i=1^n |f(x)-f(x^(i))|, and the (max) sensitivity is s(f):=max_x∈0,1^n s(f)(x).
C_3a is the largest constant such that there exist arbitrarily large sets A,B of integers such that |A+B| ≪ |A| and |A-B| ≫ |A+B|^C_3a.
For a real number γ, its irrationality exponent μ(γ) is defined by μ(γ) := infBiglc∈ℝ: Bigllvertγ-a/bBigrrvert≤ lvert brvert^-c has only finitely many solutions (a,b)∈ℤ^2Bigr. <a href="#Zud2004-def-mu">[Zud2004-def-mu]</a> We define C_7b := μbigl(Γ(1/4)bigr).
We define C_7a to be the irrationality measure of π: C_7a := sup_μ∈ℝ μ such that lvert π - p/q rvert < q^-μ for infinitely many rationals p/q. Equivalently, C_7a is the infimum of all ν such that for every ε>0 there exists q_0(ε) with |π-p/q| > 1/q^ν+ε for all integers p and all integers q ≥ q_0(ε).
Let μ be a Borel probability measure on ℝ^n with finite second moments. Its covariance matrix is Cov(μ) :=\ ∫_ℝ^n (x-m)(x-m)^mathsf T dμ(x), m:=∫_ℝ^n x dμ(x). ### Convex bodies If K⊂ℝ^n is a convex body, let λ_K be the uniform probability measure on K and abbreviate Cov(K):=Cov(λ_K).
C_20c is the KLS constant (Kannan–Lovász–Simonovits constant) for log-concave measures. It is defined as C_20c := sup_n≥ 1 ψ_n, where ψ_n is the worst-case inverse Cheeger (isoperimetric) constant among isotropic log-concave probability measures on ℝ^n.
C_11a is the smallest constant such that, for every n≥ 1 and every function f:-1,1^n → ℝ Ebigl|f(x)-Ef(x)bigr| ≤ C_11aE|∇ f|(x), where x=(x_1,…,x_n) is uniform on -1,1^n and |∇ f|(x)=Bigl(Σ_j=1^n |D_j f(x)|^2Bigr)^1/2, D_j f(x)=f(x)-f(x^(j))/2, with x^(j)=(x_1,...,x_j-1,-x_j,x_j+1,...,x_n).
C_10 is the real Grothendieck constant K_G^ℝ. It is the smallest constant C such that for every m,n ≥ 1 and every real matrix A=(a_ij) ∈ ℝ^m× n one has max_substacku_1,…,u_m, v_1,…,v_n ∈ S^∞ Σ_i=1^m Σ_j=1^n a_ij ⟨ u_i, v_j⟩ ≤ C max_ε_1,…,ε_m, δ_1,…,δ_n = ± 1 Σ_i=1^m Σ_j=1^n a_ij ε_i δ_j.
Let X be a random vector in ℝ^n with an isotropic log-concave distribution (i.e. X has a log-concave density, E X=0, and Cov(X)=Id). Since X is isotropic, E|X|^2 = n.
The Gauss–Kuzmin–Wirsing (GKW) operator acts on suitable function spaces on [0,1] by (L f)(x) = Σ_k=1^∞ 1/(x+k)^2 f (1/x+k). This is the transfer operator of the Gauss map T(x) = \1/x\, which generates the continued fraction expansion.
