長沙理工大學(xué)學(xué)術(shù)活動預(yù)告
報告承辦單位: 數(shù)學(xué)與統(tǒng)計(jì)學(xué)院
報告題目: An efficient Newton-type algorithm for a novel kind of tensor complementarity problems
報告內(nèi)容: We introduce a novel kind of structured tensor based on the concept of $P_0$ mapping, and the solvability of the corresponding tensor complementarity problem is discussed. We show that the solution set is nonempty and compact if the related tensor is an $R_0$-tensor. Especially, we also introduce a new structured tensor related to the $P$ mapping on $R^n_+$ and prove that the corresponding tensor complementarity problem has a unique solution. Based on the structure of the newly introduced tensor, we propose a smoothing Newton-type algorithm with a new one-dimensional line search rule for solving the corresponding tensor complementarity problem.
報告人姓名: 張立平
報告人所在單位: 清華大學(xué)
報告人職稱/職務(wù)及學(xué)術(shù)頭銜: 副教授、博士導(dǎo)師
報告時間: 2020年9月24日(星期四):19:30-20:30
報告方式: 線上騰訊會議,會議號ID:426-463-938
報告人簡介: 張立平,清華大學(xué)長聘副教授,博士生導(dǎo)師,研究方向最優(yōu)化理論算法及應(yīng)用,在求解互補(bǔ)與變分不等式問題、半無限規(guī)劃、張量優(yōu)化等方面取得了一些有意義的結(jié)果。已發(fā)表高質(zhì)量論文三十余篇、連續(xù)獲得多項(xiàng)國家自然科學(xué)基金資助。曾獲得教育部自然科學(xué)獎二等獎和北京市科學(xué)技術(shù)獎二等獎。