【摘要】：文章提出了一種求解帶等式與不等式約束的minimax問題的既無罰函數(shù)又無濾子的SQP算法。首先引入了ε-積極約束集,在此基礎(chǔ)上建立了兩個新的二次規(guī)劃子問題得到搜索方向,既克服了Maratos效應(yīng),又大大地減少了算法的運算量;另外給出了一種新的線性搜索步長策略,該方法既避免了罰因子的選取,又減小了計算機儲存量;在適當?shù)募僭O(shè)條件下,證明了算法的全局收斂性;初步數(shù)值實驗驗證了算法的有效性與優(yōu)越性。
[Abstract]:In this paper, a SQP algorithm without penalty function and filter for solving minimax problem with equality and inequality constraints is presented. First, the 蔚-positive constraint set is introduced, and two new quadratic programming subproblems are established to obtain the search direction, which not only overcomes the Maratos effect, but also greatly reduces the computational complexity of the algorithm. In addition, a new linear search step-size strategy is proposed, which not only avoids the choice of penalty factor, but also reduces the computer storage, and proves the global convergence of the algorithm under appropriate assumptions. The effectiveness and superiority of the algorithm are verified by preliminary numerical experiments.
【作者單位】： 太原師范學院數(shù)學系;
【基金】：國家自然科學基金(批準號:11171250)
【分類號】：O221

【相似文獻】

相關(guān)期刊論文 前1條

1 趙仲哲;關(guān)于Kuhn-Tucker定理[J];北京大學學報(自然科學);1962年04期

相關(guān)會議論文 前2條

1 鄒濤;李少遠;丁寶蒼;;基于擴張終端約束集的非線性雙模預(yù)測控制策略[A];第二十四屆中國控制會議論文集（上冊）[C];2005年

2 李桂君;;基于約束集理論(TOC)的多項目順序安排[A];現(xiàn)代工業(yè)工程與管理研討會會議論文集[C];2006年

，

本文編號：2457834