For R-D with a bunch of papers, abstract a rate-control problem to a limit condition under a restriction condition, and solve it with Lagrange to solve the idea of a cow. The research points are as follows: 1. How to simplify the complex calculation of lagrange, give a simple formula to Lagrange Multiplier. 2. How to define the formula of the Distort from a different angle. For example, the choice of Layer, Mode Decision, etc. Suddenly thinking that these articles do not seem to take into account the problem of channel errors, it seems that it is not perfect, or is my Survey doing? Also think: If you think about the ERROR ConceAlMent method of the Recevier end, Distort is not more accurate. We build the Distort model of the entire system, which combines all the three aspects of the problem, it seems better.
