We study therefore not a single operator but a set of operators, and their corresponding discrete eigenvalues, as functions of v. We devise a representation of the eigenvalue problem in which the positive-definite kinetic energy operator -Delta is replaced by a real variable s > 0. This allows the eigenvalue to be written exactly as the minimum of a semi-classical form {s + v bar{f}(s)} in which the potential-energy term is a function of s called the `kinetic potential'. The envelope method concerns potentials which can be written as a smooth transformation f = g(h) of a soluble potential h: if the transformation g has definite convexity, then the tangential potentials so generated provide energy bounds; the computation of the best such bound is carried out most effectively by means of the kinetic-potential apparatus. This same theory also allows us to discuss sums of potential terms.
In a further development, a change of variables is made which allows the semi-classical potential term to be written as the potential itself; the kinetic energy then becomes a so-called K-function. This form of the theory partially disentangles the potential and supports the goal of geometric spectral inversion, the complete reconstruction of the potential shape f(x) from a given energy function F(v).
Kinetic potentials [29][31]
Envelope theory [54]
Pure powers [45]
Sums of potential terms [26][48][65]
Symmetric potentials as mixtures of square wells [51]
Screened-Coulomb and Yukawa potentials [35][49][50][56][57]
Quartic anharmonic oscillator [33]
Coulomb-plus-linear potential [32]
Relation to the large-N approximation [36]
Singular potentials [62][63][66][69][70][72]
Log potential as limit of |x|^q as q -> 0 [54]
Dirac eigenvalues [34][37][38]
The N-body problem [25][53][60]