In applied mathematics, the Atkinson–Mingarelli theorem, named after Frederick Valentine Atkinson and A. B. Mingarelli, concerns eigenvalues of certain Sturm–Liouville differential operators.
In the simplest of formulations let p, q, w be real-valued piecewise continuous functions defined on a closed bounded real interval, I = [a, b]. The function w(x), which is sometimes denoted by r(x), is called the "weight" or "density" function. Consider the Sturm–Liouville differential equation
where y is a function of the independent variable x. In this case, y is called a solution if it is continuously differentiable on (a,b) and (p y′)(x) is piecewise continuously differentiable and y satisfies the equation (1) at all except a finite number of points in (a,b). The unknown function y is typically required to satisfy some boundary conditions at a and b.
The boundary conditions under consideration here are usually called separated boundary conditions and they are of the form:
where the { α i , β i } {\displaystyle \{\alpha _{i},\beta _{i}\}} , i = 1, 2 are real numbers. We define
Assume that p(x) has a finite number of sign changes and that the positive (resp. negative) part of the function p(x)/w(x) defined by ( w / p ) + ( x ) = max { w ( x ) / p ( x ) , 0 } {\displaystyle (w/p)_{+}(x)=\max\{w(x)/p(x),0\}} , (resp. ( w / p ) − ( x ) = max { − w ( x ) / p ( x ) , 0 } ) {\displaystyle (w/p)_{-}(x)=\max\{-w(x)/p(x),0\})} are not identically zero functions over I. Then the eigenvalue problem (1), (2)–(3) has an infinite number of real positive eigenvalues λ i + {\displaystyle {\lambda _{i}}^{+}} , 0 < λ 1 + < λ 2 + < λ 3 + < ⋯ < λ n + < ⋯ → ∞ ; {\displaystyle 0<{\lambda _{1}}^{+}<{\lambda _{2}}^{+}<{\lambda _{3}}^{+}<\cdots <{\lambda _{n}}^{+}<\cdots \to \infty ;} and an infinite number of negative eigenvalues λ i − {\displaystyle {\lambda _{i}}^{-}} , 0 > λ 1 − > λ 2 − > λ 3 − > ⋯ > λ n − > ⋯ → − ∞ ; {\displaystyle 0>{\lambda _{1}}^{-}>{\lambda _{2}}^{-}>{\lambda _{3}}^{-}>\cdots >{\lambda _{n}}^{-}>\cdots \to -\infty ;} whose spectral asymptotics are given by their solution [2] of Jörgens' Conjecture [3]: λ n + ∼ n 2 π 2 ( ∫ a b ( w / p ) + ( x ) d x ) 2 , n → ∞ , {\displaystyle {\lambda _{n}}^{+}\sim {\frac {n^{2}\pi ^{2}}{\left(\int _{a}^{b}{\sqrt {(w/p)_{+}(x)}}\,dx\right)^{2}}},\quad n\to \infty ,} and λ n − ∼ − n 2 π 2 ( ∫ a b ( w / p ) − ( x ) d x ) 2 , n → ∞ . {\displaystyle {\lambda _{n}}^{-}\sim {\frac {-n^{2}\pi ^{2}}{\left(\int _{a}^{b}{\sqrt {(w/p)_{-}(x)}}\,dx\right)^{2}}},\quad n\to \infty .}
For more information on the general theory behind (1) see the article on Sturm–Liouville theory. The stated theorem is actually valid more generally for coefficient functions 1 / p , q , w {\displaystyle 1/p,\,q,\,w} that are Lebesgue integrable over I.