Stochastic solutions for nonlinear partial differential equations

Description:  A stochastic solution of a pde is a stochastic process that started from a point in the domain, generates a boundary measure which, when integrated over the initial condition, provides a local solution of the equation. Stochastic solutions not only provide new exact solutions for nonlinear pde's but are also an efficient way to obtain local solutions without the need for a global grid. In addition, through the probabilistic domain decomposition method, they avoid the time-consuming domain communication problem in parallel computing. Both the McKean and the superprocess construction method with measures or ultradistributions are described. Date:  2019-02-20 Start Time:   14:30 Speaker:  Rui Vilela Mendes (CMAFCIO, Univ. Lisboa) Institution:  CMAFCIO, Universidade de Lisboa Place:  Room 5.5 Research Groups: -Numerical Analysis and Optimization See more:   <Main>