×

zbMATH — the first resource for mathematics

On the convergence rate of the method of successive approximations for the solution of nonlinear integral equations of Volterra type. (Russian) Zbl 0533.65092
Let f(x), K(x,s,u) be continuous in \(D=\{(x,s)\in [a,b]:\quad | u- f(a)| \leq l,\quad l>0\}; | K(x,s,u)-K(x,s,v)| \leq N(x,s)| u-v|\) in D, then for \(a\leq x\leq \min(b,a+(1-\max_{a\leq x\leq b}| f(x)-f(a)|)\cdot(\max_{D}| K(x,s,u)|)^{-1}\) there exists a unique and continuous solution of \(\phi(x)=f(x)+\int^{x}_{a}K(x,s,\phi(s))ds,\) and the sequence \(\phi_{n+1}(x)=f(x)+\int^{x}_{a}K(x,s,\phi_ n(s))ds (n=0,1,...)\) converges to it.
Moreover, \[ \max_{a\leq x\leq h}| \phi_{n+1}(x)-\phi_ n(x)| \leq \frac{1}{n!}\| \int^{x}_{a}K(x,s,f(s))ds\|_ C(\int^{h}_{a}\max_{x\in [s,h]}N(x,s)ds)^ n \] \((n=1,2,...)\). An example is given.
Reviewer: J.Albrycht
MSC:
65R20 Numerical methods for integral equations
45G10 Other nonlinear integral equations
PDF BibTeX XML Cite