Numerical solution of the obstacle problem by the penalty method. II: Time-dependent problems (Q1075740)

In part I [Computing 32, 297-306 (1984; Zbl 0528.65057)] we considered the possibility of using the penalty method to get approximations of the solution of elliptic variational inequalities. In this note we show that the same idea can be used for the treatment of parabolic problems. By coupling the penalty parameter \(\epsilon\) and the discretization parameters h and \(\Delta\) t quasi-optimal error estimates are derived in suitable norms.