Costa's corner-point conjecture for two-user Gaussian interference channels

Let (R1,R2)(R_1,R_2) be a rate pair in the capacity region of a two-user Gaussian interference channel with arbitrary positive cross-link gains a12a_{12} and a21a_{21} and power constraints P1P_1 and P2P_2. Define the single-user additive white Gaussian noise capacities

C1=12log(1+P1),C2=12log(1+P2).C_1=\frac{1}{2}\log(1+P_1),\qquad C_2=\frac{1}{2}\log(1+P_2).

Also define

R1=12log(1+a21P11+P2),R2=12log(1+a12P21+P1).R_1^*=\frac{1}{2}\log\left(1+\frac{a_{21}P_1}{1+P_2}\right),\qquad R_2^*=\frac{1}{2}\log\left(1+\frac{a_{12}P_2}{1+P_1}\right).

Costa's corner-point conjecture. For reliable communication at both receivers, the following should hold: if R2C2εR_2\geq C_2-\varepsilon for an arbitrary ε>0\varepsilon>0, then

R1R1+δ1(ε),R_1\leq R_1^*+\delta_1(\varepsilon),

where δ1(ε)0\delta_1(\varepsilon)\to 0 as ε0\varepsilon\to 0; and if R1C1εR_1\geq C_1-\varepsilon, then

R2R2+δ2(ε),R_2\leq R_2^*+\delta_2(\varepsilon),

where δ2(ε)0\delta_2(\varepsilon)\to 0 as ε0\varepsilon\to 0. The conjecture addresses the two unknown corner points of the capacity region for weak Gaussian interference channels after a gap was identified in the proof of Costa's earlier result; its validity for arbitrary positive cross-link gains remains unresolved in the supplied source.

Sources & referencesView supporting material

Primary source

Igal Sason, “On the Corner Points of the Capacity Region of a Two-User Gaussian Interference Channel”, arXiv:1306.4934 (2015).

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.