Topological Interference Management
EE Systems Seminar
We will revisit the robust principles of ignoring interference when it is weak and avoiding it when it is strong, in both cases exploring information theoretic optimality with very limited channel knowledge at the transmitters. Optimal interference avoidance will be shown to be essentially equivalent to the index coding problem which will be explored from an interference alignment perspective. Ignoring interference, i.e., treating interference as noise will be shown to optimal for the entire capacity region within a constant gap if for each user, desired signal strength is no weaker than the sum of the strengths of the strongest interference caused by the user and the strongest interference suffered by the user, with all signal strengths measured in dB scale. Several interesting open problems will also be highlighted in the talk.