Semidefinite Programming
A class of convex optimization problems in which decision variables are constrained to be positive semidefinite matrices, generalizing linear programming to matrix spaces. Solvable in polynomial time with globally optimal solutions, semidefinite programming is widely applied in combinatorial optimization and graph theory.
Timeline (last 90 days)
The semidefinite program used by SSLD to find the first color class is similar to the semidefinite program used to compute the Lovász theta number.
Semidefinite programming is a convex optimization problem that constrains decision variables to be positive semidefinite matrices; it is a convex problem solvable in polynomial time and yields a globally optimal solution.
SSLD embodies the idea of combining continuous optimization (SDP/spectral methods) with discrete heuristics: using SDP to capture the global graph structure, then using DSATUR for fast local decisions.
All Facts (3)
The semidefinite program used by SSLD to find the first color class is similar to the semidefinite program used to compute the Lovász theta number.
50%UnverifiedSemidefinite programming is a convex optimization problem that constrains decision variables to be positive semidefinite matrices; it is a convex problem solvable in polynomial time and yields a globally optimal solution.
50%UnverifiedSSLD embodies the idea of combining continuous optimization (SDP/spectral methods) with discrete heuristics: using SDP to capture the global graph structure, then using DSATUR for fast local decisions.
50%