登入選單
返回Google圖書搜尋
註釋Abstract: "Semi-Infinite programmimng, that allows for either infinitely many constraints or infinitely many variables but not both, is a natural extension of ordinary mathematical programming. There are many practical as well as theoretical problems in which the constraints depend on time or space and thus can be formulated as semi-infinite programming problems. The focus of this dissertation is on formulating and solving semi-infinite programming problems. The main results include: (1) An algorithm for solving a matrix rescaling problem formulated as a semi-infinite linear program. Sufficient conditions that guarantee finite termination are discussed and computational results are reported. (2) An algorithm for solving a matrix estimation problem equivalent to a semi-infinite quadratic program