Exponentially fitted one-step methods for the numerical solution of the scalar Riccati equation (Q1821514): Difference between revisions

An explicit exponentially fitted one-step method of Euler type is considered in order to approximate the solution of the scalar Riccati equation \[ \epsilon y'=c(x)y^ 2+d(x)y+e(x),\quad 0<x\leq X, \] with initial condition \((\epsilon >0\) is a small parameter). The method is derived using informations obtained from the constant coefficient problem. The uniform convergence (of order one) in \(\epsilon\) is proved in the case a(x)\(\neq b(x)\) and a counterexample is given for \(a(x)=b(x)\). Numerical experiments are presented making the comparison with other methods. They show that the exponential method leads to a good approximation. Moreover it is explicit and hence avoids the use of a Newton iterative procedure as for the implicit schemes.