学术讲座-Surrogate-assisted Meta-heuristic Optimization for Computationally Expensive Problems

讲座题目

Surrogate-assisted     Meta-heuristic Optimization for Computationally Expensive Problems

讲 座 人

孙超利

讲座人

职称、职务

教授

主持人

潘正祥

讲座类型

R自然科学

讲座对象

全校师生

举办时间

2018年1月12日8:00-10：00

□社会科学

举办地点

C1-206

讲

座

人

简

介

Chaoli Sun received her B.Sc. And     M.Sc. Degrees in Computer Application Technology from Hohai University,     Nanjing, Jiangsu, China, and Ph.D. in Mechanical Design and Theory from     Taiyuan University of Science and Technology, Taiyuan,Shanxi, China, in 2011.     She is a Professor in the Department of Computer Science and Technology,     Taiyuan University of Science and Technology. She was a     Research Fellow with the Department of Computer Science, University of     Surrey, working on an EU grant on SWARM_ORGAN from September 2014 to     September 2016. Her main research interests include swarm intelligence and     swarm robotics, surrogate-assisted swarm optimization, large-scale swarm     optimization, and optimization of complex mechanical systems.

She is a member of the     Evolutionary Computation Technical Committee of the IEEE Computational     Intelligence Society, an Associate Editor of Soft Computing(Springer) and an     Editorial Board Member of Complex& Intelligence System Intelligence     Society, an Associate Editor of Soft Computing. Dr. Sun has published more     than 30 papers as the first author in international journals and conferences.

讲座

主要内容

Meta-heuristic algorithms have been shown to be     powerful in optimization of engineering problems. However, a large number of     fitness evaluations are required before locating near at the global optimum,     which limits the application of meta-heuristic algorithm in solving     computational expensive problems. In this topic, a new fitness estimation     strategy, which was proposed to approximate the fitness value of individuals     based on the positional relationship between individuals, will be firstly     introduced. Then a surrogate-assisted cooperative swarm optimization     algorithm will be given subsequently for solving high-dimensional     computationally expensive problems, in which a surrogate-assisted     particle  swarm optimization(PSO)     algorithm and a surrogate-assisted social learning-based PSO(SL-PSO)     algorithm cooperatively search for the global optimum, where two algorithms     share promising solutions evaluated by the real fitness function.