特邀华南理工大学陈伟能教授来校作学术报告
发布日期:2016-09-21

报告题目:Set-based Discrete Particle Swarm Optimization and Its Applications

报告时间:923(周五) 下午 300-415

报告地点:学科3号楼S410会议室

报告人:陈伟能 教授

主持人:刘青山 教授

欢迎广大师生踊跃参加!

                   江苏省大数据分析技术重点实验室

                   江苏省气象能源利用与控制工程技术研究中心

                   江苏省大气环境与装备技术协同创新中心

                   信息与控制学院

                  2016921

报告摘要:Particle swarm optimization (PSO) is a popular swarm intelligence technique for global optimization. But since it is originally designed in continuous real vector space, most PSO variants cannot be applied to combinatorial optimization problems (COPs) directly, and the design of most existing discrete PSO approaches remains largely ad-hoc.In this talk, I will present a set-based PSO (S-PSO) framework to take advantages of PSO for solving COPs in discrete space. The main idea is to redefine the “position” and “velocity” in PSO using the concept of sets, which is a common representation in discrete space. All arithmetic operators in the velocity and position updating rules used in the original PSO are then replaced by the operators and procedures defined on crisp sets, and sets with possibilities. In this way, most of the existing PSO variants can be extended to their corresponding discrete versions with similar search behaviors. According to characteristics of the problem to be solved, alternative strategies for S-PSO in solution construction, constraint handling, usage of heuristics and scale-reduction will be further discussed. The performance of S-PSO is verified on two of the most representative COPs, the traveling salesman problem (TSP) and the multidimensional knapsack problem (MKP). In addition, I will also describe some recent applications of S-PSO on the fields of cloud workflow scheduling and search-based software engineering. Finally, raised by requirements in real world applications, some of our recent works on the extensions of S-PSO on multi-objective discrete optimization, complex multimodal optimization, mixed-variable optimization, and large-scale optimization will be introduced.

报告人简介:陈伟能,男,2016国家优秀青年科学基金获得者,华南理工大学计算机科学与工程学院教授,主要研究方向为群体智能、进化计算及其应用。发表学术论文50余篇,其中在IEEE Trans. on Evolutionary ComputationIEEE Trans, on Software EngineeringIEEE Trans. on CyberneticsIEEE Trans.刊物发表论文17篇,1篇论文入选ESI高被引论文。博士学位论文分别获得了2016IEEE CIS(计算智能学会)唯一杰出博士学位论文和2012年中国计算机学会CCF优秀博士学位论文。入选2015年“广东特支计划”科技创新青年拔尖人才,获2015英国皇家学会Newton Fund项目资助。现任中国计算机学会人工智能与模式识别专业委员会委员,多次获邀出任IEEE WCCI等领域内重要国际会议的程序委员会委员。