解无约束优化及非线性最小二乘问题的自适应信赖域算法-计算数学专业论文.docxVIP

AbstractTrust Abstract Trust region method is a kind of efficient methods to solve the general unconstrained optimization and its叩ecial situation，the nonlinear least squares problems．The choice of the trust regiou radius has important effect on the efliciency ofthe trust region method． ’{ Hei 112】proposed self-adaptive trust region“g喇thm，in which the trust region radius p is updated at a variable rate by R-function．Zhang et．a1．【26】also proposed an adaptive trust region method，in which the trust region radius d印en幽on the gradient and Hessian of the current iterate point．Numerical results show that both the methods are more e最ciont than the t￡aditional t嗽region method．The first part of this thesis proposes a newnonmonotonicself-adaptivetrust regionalgorithmwithlinesea／℃htosolvethegeneral unconstrained optimization problem，in which we combine the self-adaptive trust region method in【12】with the popul∞nonmonotone technique and use comb；,ag trust region with llne search techniques in【161．We establish the global convergence result of the new method．Numerical results show that the new nonmonotonic method can Bg％more computation than the method in【12】．The second part of this thesis puts the adaptive trust region method in 126】to floJ t0 propose an adaptive conic trust region method for nonlinear least squares problems．in which we can t|∞inexact methods to solve the trust region subproblem approxinmtely．The global and superlinear convergence results of the new algorithm established．Numerical results show that the method j8 e$cisnt． Key words．unconstrained optimization，nonlinear least squares problem，line scoffch method，trust region methed，conic model，global convergence,superlineax convergence， nonmonotone technique +y +y 981010 学位论文独创性声明 本人郑重声明； l、坚持以‘‘求实、创新”的科学精神从事研究工作． 2、本论文是我个人在导师指导下进行的研究工作和取得的研究成果． 3、本论文中除引文外，所有实验、数据和有关材料均是真实的． 4、本论文中除引文和致谢的内容外，不包含其他人或其它机构已经发 表或撰写过的研究成果． 5、其他同志对本研究所做晦贡献均已在论文中作了声明并表示了谢 意． 作者签名，圭丕拯 日 期； 壁垒：生!墨 学位论文使用授权声明 本人完全了解南京师范大学有关保留、使用学位论文的规定，学校有 权保