О некоторых тенденциях современного компьютерного моделирования

Анотація

Computer modeling (mathematical modeling and numerical simulation) has become the most powerful tool of scientific analysis in natural, technical and social sciences. However a lot of difficulties exist in computer modeling in reality, in particular, mathematical models and numerical algorithms usually belong to pure mathematical abstractions and cannot include any other kind of information beside the strictly formulated problems. Nevertheless, the modern science accumulate a huge amount of information as results of experimental investigations, analytical and approximated analytical problem solution and so on, the mentioned scientific results cannot be used in computer modeling. It is shown that this disadvantage can be overcome in the case of linear problems. For example, similar situation for linear hydrodynamic problems under small Reynolds number (Stokes flow) and high Reynolds number (potential flow) can be handled by application of boundary element method together with some iterative procedures. However there is not similar successive way in the case of nonlinear problems.