Controlling of Genetic Programming Bloat Phenomena in Rainfall-Runoff Simulation Domain

Genetic programming (GP) is an application of genetic algorithm for upgrading previously coded programs to solve problem. The main strategy of search in GP is based on genetic algorithm. Genetic programming is different in basic concepts from genetic algorithm. The difference is that genetic programming acts especially on the parse tree while genetic algorithm acts in another way, it means bit strings. These two definitions are basically different in solution methods. One of the basic applications of genetic programming is symbolic regression that is known as a route for automotive simulation. One of the main disadvantages of genetic programming is excessive growth of program without any useful effect on program itself. This phenomenon is called bloating. There are different methods for controlling of excessive growth in genetic programming. In this research the different techniques for control of bloating in symbolic regression problems were compared.Rainfall-Runoff modeling is one of the most important issues in symbolic regression of water resource ngineering, and this investigation could be effective for results improvement in Rainfall-Runoff.