Speakers
Description
The sound numerical solution of a research, high tech engineering, or economics problem involves a four step process: (i) formulation of the underlying mathematical model; (ii) design of an algorithm for solving the mathematical model in a finite number of steps; (iii) a computer code algorithm implementation, enabling floating point computations; (iv) (for expert implementation) optimization of the information flow on the different available hardware architectures (e.g., cache-based
memory hierarchy, distributed memory access, GPU accelerators).
The present lecture illustrates this four step process for the numerical solution of the Riemann integrals by automatic adaptive quadrature (AAQ), which was promoted by the rich, half century, accumulated empirical evidence as the most performant numerical method devoted to this aim.
