Two positivity preserving flux limited, second-order numerical methods for a haptotaxis model (Q2846176)

A system of partial differential equations modelling the tumor invasion through healthy tissues proposed by \textit{M.A.I. Chaplain} and \textit{A.R.A. Anderson} [Mathematical modelling of tissue invasion, Cancer modelling and simulation. Boca Raton: CRC Chapman and Hall (2003), p. 269--297] is considered. The model exhibits some complexity and due to the requirement that the solutions be non-negative, numerical methods need to be chosen with care. Standard numerical integration methods, using finite differences or finite elements, do not guarantee this and as a result of errors the computational solutions may start to oscillate and give incorrect results. In this article the aim is to develop efficient numerical algorithms with the following properties: positivity preservation; second-order convergence rate in space; optimal accuracy; correct representation of conservation; and evolution laws satisfied by the exact solutions.NEWLINENEWLINE In order to achieve this aim the authors construct two algorithms based on limiting of the flux describing the haptotaxis of cancer cells. The methods are positivity preserving second-order explicit-implicit finite difference methods based on the van Leer flux-limiter technique and adaptive mesh refinements in time. They consider mixed boundary conditions and construct second-order positivity preserving approximations also at the boundary nodes. The algorithm presented is economic time-adaptive and preserves evolution while the conservation laws satisfy the exact solution. This methodology leads to higher accuracy and to guaranteed positivity of solutions. The van Leer flux is implemented and the numerical scheme simplified by decoupling the system into three independent discrete problems. The method II shows higher precision and preserves better conservation and evolution laws whereas Method I has a better convergence rate, especially in the \(L_2\) norm.