Application Of Modern Nonlinear Optimization Algorithms On Geodetic Inversion

The study of this dissertation mainly discusses several popular modern nonlinear optimization algorithms ( Tabu Search,Simulated Annealing,Genetic Algorithms,Artificial Neural Networks,Ant Colony Optimization etc.),and the comparative advantages and disadvantages among them on geodetic inversion. Some improved methodsare given to prevent the drawback of these algorithms---slow convergence in actualuses. Then,based on Object Oriented Programming,author gives a geodetic inversion computer programme using modern nonlinear optimization algorithms.Modern nonlinear optimization algorithms have many outstanding advantages:global search,nonlinear calculate and parallel calculate etc,so they will be paid more attention on geodetic inversion. Meanwhile,they are not fitted in every situation and needed to strengthen their theoretic study,although they are general algorithms. So,we say they' re applicable if only we can get satisfying result. Gives a computer programme based on Object Oriented technology,which can make your programme clear,understandable,handled,easy-to-debug and having highly recycle rate,etc.Finally,the convergence rate of India to Eurasia subdirection is given using nonlinear optimization algorithms and GPS data. The results calculated by several algorithms are almost consistent,which indicates good stabi I i ty oT modern nonlinear optimization algorithms used on geodetic inversion. The result of geodetic inversion shows that India block moves at rates of 20.4 millimeters per year with respect to the stable Eurasia,consistent with the result comes from geology. This indicates that the rate India block moves towards Eurasia is stable.