← All problems

Alessandrini’s Problem 8.2

Let Ω⊂Rn\Omega\subset\mathbb{R}^n be a smooth bounded connected domain, let T>0T>0, and consider the diffusion problem ∂tu−∇⋅(a(x)∇u)=0\partial_t u-\nabla\cdot(a(x)\nabla u)=0 in Ω×(0,T)\Omega\times(0,T) with prescribed initial data u(⋅,0)=u0u(\cdot,0)=u_0 and prescribed boundary input u∣∂Ω×(0,T)=gu|_{\partial\Omega\times(0,T)}=g. Is there a single experiment (u0,g)(u_0,g) for which the final observation ua(⋅,T)u_a(\cdot,T) uniquely determines the scalar diffusivity a(x)a(x); equivalently, does ua(⋅,T)=ua~(⋅,T)u_a(\cdot,T)=u_{\widetilde a}(\cdot,T) imply a=a~a=\widetilde a for the relevant admissible class of diffusivities?

References

Source: arXiv

Progress summary

Refreshed
Open · progress reported

A new study gives a positive answer only for tightly controlled material classes, so the unrestricted inverse problem remains open.

Alessandrini’s Problem 8.2 asks whether a coefficient can be recovered from a single final observation in a diffusion experiment. The new result addresses this question under prescribed uniformly elliptic, C1,αC^{1,\alpha}-bounded coefficients with known boundary values, rather than under unrestricted assumptions.

October 2026 restricted-class result

Ching-Lung Lin and Yi-Hsuan Lin claim global uniqueness and Lipschitz stability at every positive final time from one constant initial datum and spatially constant exponential boundary input. They also identify an optimal Sobolev-scale norm for the measured data. This is substantive progress for the stated coefficient classes, but the claim is unverified and does not settle the unrestricted problem.

Current status (as of October 2026): Recovery is claimed for broad but explicitly restricted coefficient classes; the unrestricted form of Alessandrini’s Problem 8.2 remains open.

Sources

Solutions 0

No solutions have been posted yet.