老师介绍：Stephen Boyd， 这是他的书网站Convex Optimization http://web.stanford.edu/~boyd/cvxbook/，有pdf和课程。

Kown the big picture 模块化,了解概念，会应用到实际问题，就挺好，很多数学很难的。

Many questions have no analytical solution

Convex not need know the gridient

Concave(convex) none thing

Convex(concave) have rules

Many old method need to the goal function is differentiable for first or more than second times,

But new method for convex problems don‘t.

Concave(convex) is don‘t know