We deal with the numerical solution of elliptic not necessarily self-adjoint problems. We derive a posteriori upper bound based on the flux reconstruction that can be directly and cheaply evaluated from the original fluxes and we show for one-dimensional problems that local efficiency of the resulting a posteriori error estimators depends on p 1/2 only, where p is the discretization polynomial degree. The theoretical results are verified by numerical experiments.

中文翻译：

---

关于通量重建的多项式鲁棒性

我们处理椭圆不一定是自伴随问题的数值解。我们基于通量重建推导出后验上限，可以从原始通量中直接且廉价地评估该上限，并且我们针对一维问题表明，所得后验误差估计量的局部效率仅取决于 p 1/2，其中 p是离散化多项式次数。通过数值实验验证了理论结果。