Performance of three solvers(WDL-PDE,R-WDL-PDE and DRNN-WS-PDE ) on Allen-Cahn equation

These data (that is, the solution, the loss，the CPU time and their respective standard deviations) describe the performances of three solvers, i.e., WDL-PDE, R-WDL-PDE and DRNN-WS-PDE, on Allen-Cahn equation. The run times is set to 100. All experiments are carried out independently. Numerical results output from three solvers are listed.To ensure good comparability, the number of logic operation neural units contained in DRNN-WS-PDE solver and WDL-PDE solver are kept the same; and the same random seed is used to generate the path in experiments. Besides, all experiments are conducted on the same platform, i.e., Google colab equipped with 12.7 GBP virtual memory (RAM) and Python 3.8’s TensorFlow 2.11.0 library.