Piecewise static policies for two-stage adjustable robust linear optimization