Reinforcement Learning for Adaptive Mesh Refinement

Burgers equation

IPN trained on 1-bump fixed initial condition

Advection of a bump function

Train budget 20 | Test budget 50

Train budget 20 | Test budget 50

Train budget 20 | Test budget 50

Train budget 20 | Test budget 100

Advection of a circle function

Train budget 20 | Test budget 50

Train budget 20 | Test budget 50

Static-trained policies generalize to advection

Train: static budget 10

Test: advection budget 50

Train: static budget 10

Test: advection budget 50

Train: static budget 10

Test: advection budget 50

Page updated

Google Sites

Report abuse