An Improved GEP-GA Algorithm and Its Application

GEP is a powerful tool for automatically function modeling. However, the classical GEP have some appearances such as lack of learning mechanism, search blindly, lack of diversity, prone to precocity when dealing with complicate problems. In light of these limitations the Improved GEP-GA Algorithm introduces the uniform initial population strategy, the adaptive mutation, the variation of population size strategy based on stagnant generations, and optimizes the coefficient of model by GA after the work of GEP. Then it proved that the Improved GEP-GA Algorithm is more effective than other similar algorithms in modeling and forecast though some experiments. It will make the algorithm hard to get into the local trap, and improve the fitting efficiency and forecasting accuracy of mining. The result of that applied the improved GEP-GA algorithm to the relay’s parameter design shows that it will save time in calculating the electromagnetic suction force and improve the computational efficiency in a large degree. It has a broad apace for development and application in relay parameters design.