Statistical Analysis of Combined Simplex Search

The statistical analysis of combined simplex search was made using a multiple Markov chain to describe the search process in noisy surrounding or aim drift conditions. Investigations showed that if an optimized object is impacted by non-controlled but observed (measured) factors, evaluation of their parameters by using algorithms of combined simplex search capacitates the improvement of the efficiency of the search for extremum. During the climbing stage it is invariant to the measured accidental noise, depends only on the drift speed, has no instable mode. These results can be used to found new effective algorithms of simplex search.