Symbolic Regression Based on GEP and Its Application in Predicting Amount of Gas Emitted from Coal Face

摘要

Gene Express Programming (GEP) can discover relationships among observed data and express them mathematically. Moreover, GEP also inherits the advantage of easy genetic manipulation from Genetic Algorithms (Gas) and therefore is more efficient than Genetic Programming (GP). This paper presents a new Symbolic regression Algorithm Based on GEP, and uses it to compute the Amount of Gas Emitted from Coal Face. Two numerical examples are presented to illustrate for finding the Regression equation. Results show that the model discovered by gene expression programming is much more accurate and stable than the one discovered by genetic programming or linear regression.