LP τ search with averaging makes it possible to obtain an approximate solution for the problem of searching the region of efficiency in the multidimensional parameter space such that values of quality indices in this region will be better than those of other region of parameter values (the rational solution of the opti mization problem). In order to describe the region of efficiency, it is proposed to use identification algo rithms. We present an example of a simulation model for which the use of recurrent identification algorithms for approximating the region of efficiency of a dynamic stochastic object is analyzed.