Parallel computations in the problem of analysis of a mechanical system movement in a thermoviscoelastic medium

This paper presents a technique of application of parallel computing for approximate probabilistic analysis of a behavior of a dynamic mechanical system with taking into account thermoviscoelasticity of medium. To study this system described by linear stochastic integro-differential equations, a modification of the iterative approximation method of Green's functions and the semi-implicit method of Euler^Maruyama method combined with statistical modeling were used. We present a general description of the model, the scheme of usage of the well-known computer algebra Mathematica package for the preparation of calculations, breefly describe algorithms and methodology of application of parallel computing, present some results of experiments on a computer with four-core CPU, conclusions and recommendations.