时间：2021年10月28日周四 9:00-11:00
报告人：陆芷、周亦威
地点：经管大楼A楼四楼第二会议室报告厅
报告一
题目：Optimization Approachesfor Minimum Conductance Graph Partitioning
内容简介： The minimum conductance graphpartitioning problem (MC-GPP) is an important NP-hard combinatorialoptimization problem with numerous practical applications in various areas suchas community detection, bioinformatics, and computer vision. Due to its high computationalcomplexity, heuristic and metaheuristic approaches constitute a highly usefultool for approximating this challenging problem. This study is devoted todeveloping effective metaheuristic algorithms for the MC-GPP. Extensivecomputational experiments and comparisons on large and massive benchmarkinstances (with more than 23 million vertices) demonstrate that the proposed algorithms compete very favorably with state-of-the-art algorithms in theliterature.
报告人介绍：陆芷 博士毕业于法国昂热大学 (Universitéd'Angers)，专业为计算机科学。现任上海理工大学管理学院信息管理与信息系统系讲师。研究方向包括运筹决策、图与组合优化问题，启发式算法、优化理论与方法等。研究成果主要发表在IEEETransactions on Cybernetics, Computers & Operations Research和Information Sciences等期刊上。目前主持国家自然科学基金青年项目一项。
报告二
题目：Investigating travel flow differences between peak hours with massivemobile phone data using a spatial model with endogenous weight matrix
内容简介： Traffic congestion is an important topic intransportation research. Urban travel flow at peak hours has its own patternsand is influenced by various factors. To account for possible endogeneity inspatial effects among traffic analysis zones (TAZs), this paper establishes aspatial model with an endogenous weight matrix to investigate the travel flowdifferences between morning peak and evening peak on both weekday and weekendbased on mobile phone data in Hangzhou, China. The results indicate that thelogarithm sum of supermarkets, shopping malls, and parks has positive impactson travel flow differences on both weekday and weekend. Moreover, the endogenousweight matrix on both weekday and weekend are successfully estimated andcompared. The above empirical analysis reveals the mechanism of spatialinfluence with endogeneity, deepens understanding of urban travel flow betweenpeak hours, and facilitates urban planning and policy making.
报告人介绍：周亦威博士毕业于美国伦斯勒理工学院 (Rensselaer Polytechnic Institute)，专业为交通运输工程。现任上海理工大学管理学院信息管理与信息系统系讲师。研究方向包括空间计量模型、交通出行行为分析、自动驾驶测试评价等。研究成果主要发表在Transportation Research Part A, Part B, Part C等期刊上。目前主持上海市政府发展研究中心决策咨询课题一项。
