The Setting
On an interval where the order parameter is constant, , write , and for the scaled spatial derivatives of the value function . The quartic form (1) has a nonpositive middle term where , and two nonnegative ones. The quartic form (1) is coercive on average against a probability density on the line if for some constant .
The library states two bounds of this kind for real functions on the line.
The Polynomial and Its Flow Derivative
The first bound rests on a cubic polynomial in five real variables , , , and , whose coefficients are built from the rational constants (2)
From these constants, the polynomials in and in are the quadratic forms (3) and the polynomials and in are the cubic forms (4)
In each variable, the partial derivative of is a quadratic form: (5)
The flow derivative at real numbers is the derivative of along the vector field , with each partial of (5) taken at ✦, and the weighted derivative at the same point adds to it: (6)✦
Tilted Gaussian Weights and the Backward Shape
A bounded smooth function is a function on of class whose derivatives of every order are bounded: for each there is a constant with for all .✦ A regular datum is a function on of class whose derivative is a bounded smooth function.✦
The Burgers rate of a function of time and space is .✦ The backward Wronskian of is the Wronskian of and its Burgers rate in the space variable.✦ A function on has the backward shape if it is odd, has everywhere, and at every satisfies the three inequalities of (7). (7)✦
The coercive polynomial of a real function on is the form (1) with , and replaced by the certificates of , its first three derivatives , and : (8)✦ and the quartic moment of is the function ✦.
For a real number and a real function on , the tilted Gaussian density is the unnormalized weight on given by (9)✦
The Claim
Theorem 1. (a) Let be a nonnegative function and , , , , , , , real functions on with , and everywhere, and let and stand for (4) and (6) evaluated along and . Suppose that the flux has derivative at every point and tends to at and at , and that , and are integrable on . Then (10)✦
(b) Let be a bounded smooth function with the backward shape, and let be an odd regular datum with everywhere and with and on . Let , and let be an even Lipschitz function on with at every . Then (11)✦
In the paper, the result reads: if the variables are related as in Lemma 3.4, with its derivative rules, with , and , and with their quotients extending across zero, and if is a probability density with for which every term of the identity of Lemma 4.1 is integrable and at , then .
What Is Checked
Part (a) of the theorem is stated over the constants (2), the polynomials , , and , the partial derivatives of , its flow derivative and weighted derivative , and the hypotheses on the weight and on ; part (b) is stated over bounded smooth functions and regular data, the backward shape with the Burgers rate and backward Wronskian it is defined by, the coercive polynomial of built from its certificates , , , its quartic moment, and the tilted Gaussian density .