Many uncontrollable factors cause random fluctuations in the properties of an oil pipeline. After a brief statistical analysis of the leading parameters used to identify the phenomenon, this article describes an optimization algorithm for minimizing energy consumption in pumping stations. The proposed algorithm is based on a linear regression model. Several very flexible approaches to multivariable identification are examined.