A linearly convergent doubly stochastic Gauss-Seidel algorithm for solving linear equations and a certain class of over-parameterized optimization problems.
Meisam RazaviyaynMingyi HongNavid ReyhanianZhi-Quan LuoPublished in: Math. Program. (2019)
Keyphrases
- linear equations
- gauss seidel method
- cost function
- dynamic programming
- objective function
- particle swarm optimization
- computational complexity
- np hard
- probabilistic model
- optimization problems
- expectation maximization
- k means
- evolutionary algorithm
- simulated annealing
- optimization algorithm
- matching algorithm
- convergence rate
- similarity measure