Method of fictitious domains and homotopy as a new alternative to multidimensional partial differential equations in domains of any shape (Q2203100)

The authors investigate a newly developed coupled method of fictitious domains and homotopy for solving multidimensional partial differential equations in domains of any shape. The complexity of dealing with the geometries of these domains has caused a real challenge associated with solving such equations; therefore, there is a great need for an alternative efficient method to overcome this challenge. On the one hand, the authors first apply the method of fictitious domains to the studied model problem, where this problem can be approximated by another problem in a canonical domain known as parallelepiped. On the other hand, the homotopy method is applied, so the studied problem in canonical domain is transferred to a more complicated form of this problem in a domain of any shape using an artificially proposed parameter with subsequent algorithm application as mentioned by the authors. Various cases are discussed using the following boundary value problem (model problem): Given \(\Omega\) as an arbitrary domain, then we have the following model problem: \[ Lu(x) \equiv -\sum_{i,j=1}^{n} \frac{\partial}{\partial x_{i}} \left(a_{ij}(x) \frac{\partial u}{\partial x_{j}} \right)+c(x)u = f(x), \] where \(x=(x_{1},\dots,x_{n})\in \Omega\). This problem is subjected to the following conditions: \(u(x)=0\), where \(x \in \partial \Omega\); \(a_{ij}(x)=a_{ji}(x)\), where \(c(x) \geq 0\); \(\inf_{x \in \Omega} \sum_{i,j=1}^{n} a_{ij}(x)\xi_{i}\xi_{j} \geq \mu \sum_{i=1}^{n} \xi^{2}_{i}\), where \(\mu\) is a positive constant which is an independent of an arbitrary vector \(\xi=(\xi_{1},\dots,\xi_{n})\). The authors obtain an exponentially convergent sequence of solutions to the solution of the studied boundary value problem in parallelepiped. Finally, the proposed method is discussed in a very simple way and also applied to various problems and cases such as spectral problems, and polygonal domains in 2D and in 1D.