Method of successive approximations for solving integral equations of the theory of risk processes (Q2574220)

It is known that the equation \(\xi_1=u+ct-S_t\) defines the evolution of the capital \(\xi_1\) of an insurance company, where \(t\) is time, \(u\) is the initial reserve of the insurance company, \(c\) is the intensity of inflow of premiums, \(S_t\) are aggregated payments by time \(t\) which are given by \(S_t=\sum_{k=1}^{N_t} z_k\), \(z_k\) are equally distributed independent random variables with the distribution function \(F(z)\) and average value \(\mu\), \(F(z)=0\) for \(z\leq0\) and \(N_t\) is the number of payments by the time \(t\) (a Poisson process with intensity \(\alpha\)). It is also known that the function of non-bankruptcy probability \(\phi(u)=Pr\{\xi_t\geq0\;\forall t\geq0\}\) satisfies, for the initial capital \(u\), the following integral equation \[ \phi(u)=\int_0^\infty \alpha e^{-\alpha t}\int_0^{u+ct} \phi(u+ct-z)dF(z)\,dt,\quad \phi(u)=0\;\text{ for }u<0\quad\text{and }\phi(\infty)=1. \] This can be put in the form \[ \phi(u)=1-\frac{\alpha\mu}{c}+\frac{\alpha}{c} \int_0^u \phi(u-z)(1-F(z))\,dz.\tag{\(*\)} \] In his earlier work [Theory of optimal solutions, No. 2, V. M. Glushkov Inst. Cybern., NAS Ukr., 10--18 (2003)], the author has obtained the solution of (\(*\)) by the Picard method of successive approximations \[ \phi^{k+1}(u)=1-\frac{\alpha\mu}{c}+\frac{\alpha}{c} \int_0^u \phi^k(u-z)(1-F(z))\,dz,\quad k=0,1,\dots. \] In the present paper, he generalizes his previous work to the general (non-Volterra) integral equations of the theory of risk processes, including risk processes in Markovian and semi-Markovian media.