← All problems

Concavity conjecture for the logarithm of a matrix permanent

17 years old · traced to

Let A{\bf A} be an M×NM\times N matrix with non-negative elements. For λ=(λ1,…,λN)\boldsymbol\lambda=(\lambda_1,\ldots,\lambda_N), let

f(λ)=log⁡Per⁡(Adiag⁡(λ)).f(\boldsymbol\lambda)=\log \operatorname{Per}({\bf A}\operatorname{diag}(\boldsymbol\lambda)).

Define

DN={λ∣Per⁡(Adiag⁡(λ))>0, λi≥0, 1≤i≤N}.\mathcal{D}^{N}=\{\boldsymbol\lambda\mid \operatorname{Per}({\bf A}\operatorname{diag}(\boldsymbol\lambda))>0,\ \lambda_i\geq 0,\ 1\leq i\leq N\}.

Concavity conjecture. The function f(λ)f(\boldsymbol\lambda) is concave on DN\mathcal{D}^{N}.

This conjecture concerns the concavity of the logarithm of a permanent under non-negative diagonal scaling and is intended for the optimal power allocation problem in jointly correlated MIMO channels.

References

Status

  • Solved: Leonid Gurvits. A proof of the log-concavity conjecture related to the computation of the ergodic capacity of MIMO channels (arXiv, 2009) https://arxiv.org/abs/0911.0696

Source: Xiqi Gao, Bin Jiang, Xiao Li, Alex B. Gershman and Matthew R. McKay, “Statistical Eigenmode Transmission over Jointly-Correlated MIMO Channels”, arXiv:0903.1952 (2009).

Progress summary

Never refreshed

Nothing recorded yet. Refresh searches the literature and the public web for attempts on this problem, and writes the first summary here.

Solutions 0

No solutions have been posted yet.