Large-Scale Scientific Computing

Volume 5910 of the series Lecture Notes in Computer Science pp 645-652

Additive Operator Decomposition and Optimization–Based Reconnection with Applications

  • Pavel BochevAffiliated withApplied Mathematics and Applications
  • , Denis RidzalAffiliated withSandia National Laboratories, Optimization and Uncertainty Quantification


We develop an optimization-based approach for additive decomposition and reconnection of algebraic problems arising from discretizations of partial differential equations (PDEs). Application to a scalar convection–diffusion PDE illustrates the new approach. In particular, we derive a robust iterative solver for convection–dominated problems using standard multilevel solvers for the Poisson equation.