Optimization Inverse Spectral Problems for Sturm–Liouville Operators with Potentials in $L^p(0,1)$

We study an optimization inverse spectral problem (OISP) for a one-dimensional Sturm–Liouville operator on a finite interval. The problem is to find a potential $V \in L^p(0,1)$ that realizes a prescribed finite set $λ_1^* < λ_2^* < \dots < λ_m^*$ and has the smallest $L^p(0,1)$-distance from an a priori given potential $V_0 \in L^p(0,1)$. We formulate and prove results on the regularity of the isospectral manifold, the existence of an OISP solution, and an explicit analytic representation of the optimal potential by means of a power map applied to a linear combination of squared eigenfunctions. Using a unified differential-algebraic approach to squared eigenfunctions, we show that whenever these squares are linearly dependent, the corresponding potential can be expressed in terms of them and is an analytic function. For $p=2$, we prove that nondegenerate stationary points are isolated and that the number of global minima is finite provided that each of them is nondegenerate. For every $1<p<\infty$, we establish uniqueness for the first eigenvalue when $λ_1^*\geλ_1(V_0)$ and generic global uniqueness in the sense of Baire category in $L^p(0,1)$. A separate section provides examples with at least two global minima for $m=1,2$ in the presence of a sufficiently high finite symmetric barrier. We describe an iteration with frozen eigenfunctions and a local Newton–KKT method with continuation in the spectral data, and state explicitly the limits of their guarantees. The final part discusses applications of OISP theory to nonlinear boundary value problems and to applied problems of diagnostics and optimal spectral design of acoustic and electromagnetic resonators.