Picard's iterative method for nonlinear advection-reaction-diffusion equations (Q1036549)

A system of nonlinear partial differential equations of the type \[ u_t(t,x) =F(t,x,u(t,x),\nabla u(t,x),\Delta u(t,x)), \;\;t_0<t<T<\infty , x\in \Omega \subset \Re^m , \] \[ u(t_0,x) =u^{(0)}(x), \] with smooth in time exact solution is investigated. This type of equations governs many types of real physical problems arising in heat an mass transfer, combustion, population dynamics, ecology, chemotaxis, biology etc. Under the assumption that an unique exact solution exists, the problem can be rewritten in a Volterra integral equation form and this is the main idea also for proposed numerical approximation of the exact solution. The presented iterative method is based on Picard's iteration. The convergence of the numerical solution is based on the introduction of a complete metric space in which a contractive mapping is constructed and Banach's fixed point theorem is used. A variational iteration method is derived by means of simple integral identities without making use at all of Lagrange multipliers and constrained variations. This technique provides exactly the same sequence of iterates as Picard's method if the initial guess is equal to the initial condition. The convergence of the numerical solution to the exact one is proved under some hypothesis on the function F (boundedness in functional space, Lipschitz in last three variables) and is local in time. Moreover some generalization of the iterative procedure governs except advection-reaction-diffusion equations also another type of problems is presented too. Finally, several numerical examples, that confirm advantages and shortcomings of the iterative method are shown.