Gaussian Channel continuous alphet channel: input XiX_iXi, noise ZiZ_iZi, output YiY_iYi Gaussian channel: Yi=Xi+Zi,Zi∼N(0,N)Y_i=X_i+Z_i,Z_i\sim\mathcal{N}(0,N)Yi=Xi+Zi,Zi∼N(0,N) Energy Constraint: 1n∑i=1nxi2≤P\frac{1}{n}\sum_{i=1}^n x_i^2\leq Pn1∑i=1nxi2≤P C=maxf(x):EX2≤PI(X;Y)C=\max_{f(x):EX^2\leq P}I(X;Y)C=maxf(x):EX2≤PI(X;Y) Gaussian channel C=12log(1+PN)C=\frac{1}{2}\log (1+\frac{P}{N})C=21log(1+NP), maximum attained when X∼N(0,P)X\sim\mathcal{N}(0,P)X∼N(0,P) N=EZ2N=EZ^2N=EZ2 信噪比:PN\frac{P}{N}NP Guassian Noise is worest noise