In this thesis,we study three topics about weighted mean inequalities new combinatorial mean inequality and singular values of matrices.Our main results are divided into three parts.First,we prove following a new mean inequality,i.e.,(?)where ai,bi are positive real numbers(i=1,2,...,n)and p,q are arbitrary real numbers by using the elementary inequality(?) here fi:E(?)R?R,(i=1,2,...,n)and constructing auxiliary functions.Using this inequality,we deduce the harmonic-geometric-arithmetic square mean inequality.As an application,a new characterization of unitary matrix is given by using the formula and numerical radius.Second,a new combinatorial arithmetic-geometric mean inequality is established by using multivariate polynomial theorem and we solve an No.12066 open problem about determinant of positive definite matrix in the American Mathematical Monthly.At last,we obtain a characterization of pair matrices such that sj(A)?sj(B),j=1,…,n where s,(T)denotes the j-th largest singular values of T.Using this characterization,we give a new proof of Zhan's result related to singular value inequality of differences of positive semidefinite matrices. |