The Setting
For a natural number and a bounded interval , take an orthonormal family of real functions on and positive weights, as arise in the spectral theory of one-dimensional Schrödinger operators. Balancing a regularized matrix field on real symmetric matrices against a rank-one outer product scaled by a lower triangular matrix yields a singularly perturbed matrix flow with parameter . The question is how to join opposite diagonal matrices at and by a symmetric matrix path along this flow while keeping the path and the lower triangular factor controlled uniformly as becomes small.
Matrix Spaces and the Regularized Field
Given a natural number , the symmetric matrix space is the space of real symmetric matrices.✦ For a natural number , a lower triangular matrix is an real matrix such that whenever .✦ For a natural number , the lower triangular matrix space is the space of all lower triangular real matrices.✦
To each vector (for a natural number ) corresponds the diagonal symmetric matrix , the diagonal matrix with -th diagonal entry .✦ Any vector (for a natural number ) determines the symmetric outer product , the symmetric matrix with entry .✦
For a natural number , let be the identity matrix, for a vector let be the diagonal matrix with diagonal entries , and for an real matrix and , let the matrix pencil be the real matrix (1)✦ For real numbers , , and , the regularized scalar primitive , with , is the real number defined by (2)✦ For a real parameter , the field normalization constant is the real number (3)✦
For a natural number , parameters , a vector , and an real matrix , with evaluated by the continuous functional calculus using (1), (2), and (3), the truncated matrix field is the real matrix (4)✦ In terms of the truncated matrix field (4), the matrix field for , a natural number , , and an real matrix is the real matrix (5)✦ For a natural number , parameters and , and a vector , the symmetric matrix field is the map from to sending each to the matrix from (5).✦
For any natural number , a vector , and an real matrix , the linearized endpoint map inverse is the real matrix (6)✦
The Claim
Theorem 1. Let be a natural number, let satisfy and , and let satisfy . Let satisfy for every , and let be continuous on such that, for all and , equals when and when . Then for each , there exist a lower triangular matrix and a path , differentiable within at every with derivative , satisfying (7), where is bounded entrywise by a constant independent of . (7)✦
In the paper, the result reads: for a bounded interval , a function satisfying , and with every , assuming and have the regularity and properties (5.1) through (5.3), for every there exist a real lower triangular matrix and a path such that, with , , , , and for , where for fixed , , and the matrices can be chosen bounded uniformly for . The theorem above answers that statement for the interval with , , and , replacing by continuity of on and the abstract pair by , while restricting to and stating (7) via differentiability within , an entrywise bound on , and the matrix and vector norms in place of the operator norm.
What Is Checked
The connecting path claim (7) is stated over a natural number , real parameters , , and , a positive vector , and a continuous orthonormal function on , using the lower triangular matrix space , the symmetric matrix space , the diagonal matrix , the symmetric outer product , and the symmetric matrix field built from (1), (2), (3), (4), and (5), alongside the linearized endpoint map inverse (6) defined on vectors and real matrices .
Suppose and are such that is continuous on , differentiable at every with derivative , satisfies , and obeys whenever and ; then the upper barrier bound (8) holds: (8)✦
If is a natural number, is an lower triangular real matrix, and satisfies for every and , then the Gram diagonalization (9) gives (9)✦