Argonne’s Mathematics and Computer Science division is developing models, theory, algorithms, and scalable implementations to build a rigorous mathematical foundation for addressing scientific and engineering challenges.
Focusing on the development and numerical analysis of scalable algorithms for solving systems involving complex multiphysics, multiscale phenomena through projects such as A mixed-integer PDE-constrained optimization formulation for constructing electromagnetic cloaks with multiple materials and PETSc/TAO.
Conducting strong multidisciplinary research activities encompassing linear algebra, adjoint-based techniques, and uncertainty quantification for stochastic systems, through projects such as PETScv TSAdjoint and Optimal checkpointing for adjoint multistage time-stepping schemes.
Developing new mathematical formulations, underlying theory, and methods for solving optimization problems arising in control, data assimilation, experimental design, inverse problems, and machine learning, through projects such as Data-Driven Optimization under Uncertainty, Optimizing Stochastic Grid Dynamics at Exascale, and MINOTAUR.