... for problem with equality constrains only and obtains an approximate solution, which may not satisfy inequality constraints. Then Dikin Interior Point Method starts from the approximate solution and finds an optimal one. For problems of high dimension sparse matrix methods, namely Conjugate Gradient and Cholesky method with nested dissection, are applied. Since Dikin Interior Point Method works much slower then Newton Method on the matrices of big size, such approach allows us to obtain good starting ...
Теги: convex optimization , interior point method , large-scale optimization , newton method , pipe network , sparse matrix , constraint theory , linear programming , newton-raphson method , optimization , quadratic programming , approximate solution , inequality constrain
