第一场：
Augmented Lagrangian Methods of Multipliers for VLSI Global Placement

Abstract: We proposed several augmented Lagrangian methods of multipliers for solving the large scale constrained nonlinear programming raised in the flied of the very large-scale integration (VLSI) placement problem, and established some convergence properties of the proposed methods. Experimental results on the International Symposium on Physical Design (ISPD) benchmarks, compared with some state-of-the-art methods, show that the global placement methods resulted by the proposed methods are very efficient.

讲座时间：12月13日下午4:00 
讲座地点：博学楼学院四楼会议室（416-1）

第二场：
Trace Lasso Regularization for Adaptive Sparse Canonical Correlation Analysis with Applications

 Abstract: Canonical correlation analysis (CCA for short) describes the relationship between two sets of variables by finding some linear combinations of these variables that maximizing the correlation coefficient. However, in high-dimensional settings where the number of variables exceeds the sample size, or in the case of that the variables are highly correlated, the traditional CCA is no longer appropriate. In this paper,  an adaptive sparse version of CCA (adaptive SCCA for short) is proposed by  using the trace Lasso and Lasso regularization. The proposed adaptive SCCA reduces the  instability of the estimator if the covariates are highly correlated,  thus improves its interpretation.  The adaptive SCCA model is further reformulated to an orthogonality constrained optimization problem, and an effective splitting method is then proposed for solving the resulting  optimization problem.  The performance of the  SCCA model is compared with the other sparse CCA techniques in different simulation settings,  and its validity is also illustrated on the real genomic data sets.

讲座时间：12月14日上午9:00 
讲座地点：博学楼学院四楼会议室（416-1）

简介：彭拯博士， 福州大学数学与计算机科学学院教授，博士生导师。主要从事数学优化及其应用领域的算法设计与分析研究，发表论文近40篇，主持和参加国家自然科学基金等科研项目10余项。目前担任福建省运筹学会秘书长，运筹学与控制论福建省高校重点实验室副主任等学术职位。