Why it matters

Convex optimization is workhorse. Understanding shapes ML + OR.

Advertisement

The architecture

Objective convex + feasible set convex.

Any local min is global.

Polynomial algorithms.

Convex optimization structureConvex objectivef convexConvex feasible setpolyhedral / ball / etc.SolveIPM / gradient / etc.Boyd + Vandenberghe textbook standard; CVXPY + solvers cover practice
Convex optimization.
Advertisement

How it works end to end

Boyd + Vandenberghe textbook.

CVXPY + solvers.

SOCP + SDP + LP + QP hierarchy.

Duality theory.