The parallel solution for the task of numerical modeling of pollution expansion in the air space of cities and big plants

The numerical modeling of pollution expansion in parallel mode is considered. The mathematical model of gaseous and dusty pollution expansion takes into attention the heat, turbulence, chemical reactions and other factors. This model is very homogeneous that allows to perform an effective parallel solution. The principles of selection of the parallelizing method for considered task are described. The new algorithm of distributed load balancing is introduced. This algorithm has a minimal number of data sends/receives. The effectiveness of parallel solution is investigated. The simulation results are compared with real data.