The Comparison Of Genetic Algorithms And Genetic Programming

Genetic algorithm originated from Darwin's evolution theory and Mendel's genetic the-ory, and reflected the "survival of the fittest" principle in nature. Guided by the individual's fitness, genetic algorithm retained the best acclimation individual by the incessant evolution, and finally gain the optimal solution. The implementation of this process simulates the evo-lution in nature, and also undergos the selection process,the crossover process,the mutation process and so on.The genetic algorithms's operation objects are fixed-length strings, while the genetic programming's are hierarchical computer programs with dynamic variable size and struc-ture. Genetic programming inherits the highly parallelism and searching function of genetic algorithm, what's more, it featured with hierarchical and dynamic structure.As a viable new technology, a large number of scholars are working on the research of genetic algorithm and genetic programming, including theory research and application re-search. But genetic algorithm still has a lot of disadvantages, mainly about slow convergence and premature phenomena, so we also need to do more deeply study on it; as an improvement of genetic algorithm, the study on genetic programming is particularly important.The main woke of this paper are the following:Firstly, summarizing the basic theories and operation techniques of genetic algorithm. In this paper, the operation process of genetic algorithm is described in detail, including cod-ing process, the selection process, the crossover process, the mutation process and so on; then summarizing the techniques used for each process, such as:coding methods, the trans-formation of fitness, selection methods and so on; and the analyzing the influence parameters of genetic algorithm; finally,the proof of schema theorem and implicit parallelism theorem which are the mathematical foundation of genetic algorithm are given.Secondly, summarizing the basic theories and operation techniques of genetic program-ming and introducing several methods to improve genetic programming. Combining with the learning of genetic algorithm, summarizing and generalizing the theory of genetic program-ming, focusing on coding, the methods and techniques of crossover and mutation, which are different from genetic algorithm's, and introducing several improved methods for its short- comings, such as too complex individuals, slow convergence, a single group and so on.Thirdly, comparing genetic algorithm with genetic programming comprehensively. Mak-ing a detailed comparison of genetic algorithm and genetic programming on the theoretical basis, the essential characteristics, the operation methods and the scope of application, re-vealing the essential relations and differences between genetic algorithm and genetic pro-gramming, by comparing the two methods recognizing the connotation and characteristics of them, and giving a brief guiding principle of the two methods where and how they are applied.Fourthly, for the prediction of settlement in the dam, using the principles of genetic programming and curve fitting method in order to establish prediction model which influence factor is time, and comparing with the hyperbolic regression model and genetic algorithm model, then we obtained that genetic programming has a higher accuracy.In theory, doing a thorough research of genetic algorithm and genetic programming, especially comparing one with each other. In practice, using C++language program genetic programming which make it have been effectively applied in the prediction model.