【直播】【青年科学半月谈】混合整数规划基础及在工程科学中的应用

直播信息

报告题目

混合整数规划基础及在工程科学中的应用

报告人（单位）

Liding Xu（École Polytechnique (LIX)）

报告时间

2022年6月9日（周四）13:00

主办方

蔻享学术

直播二维码

直播海报

报告人介绍

Liding Xu is a second-year Ph.D. student in the computer science lab at École Polytechnique (LIX), where he is part of the OptimiX group. His research interests are in the area of mixed-integer linear and nonlinear programming. Currently, he is working on the following topics: decomposition methods, cutting plane methods, applications in wireless networks, and urban mobility problems. Most of his research is based on open-source software.

报告摘要

Mixed-integer nonlinear programming (MINLP) problems are a family of mathematical programming problems that contain integer decision variables and nonlinear constraints/objective. Current off-the-shelf global solvers can solve some MINLP problems to global  ε-optimality. In this talk, I will give a short introduction to MINLP and my recent work on submodular binpacking problems. Computationally tractable relaxation is crucial for MINLP. Tremendous works in the literature build relaxations for various MINLP problems. Submodular binpacking is used to model the classical binpacking problems under uncertainty. We build a Dantzig-Wolfe relaxation for the submodular binpacking, and propose several algorithms to solve it. The computational result shows that our algorithms are faster than the state-of-art commercial solver.

扩展阅读

 

编辑：王亚琨

蔻享学术平台，国内领先的一站式科学资源共享平台，依托国内外一流科研院所、高等院校和企业的科研力量，聚焦前沿科学，以优化科研创新环境、传播和服务科学、促进学科交叉融合为宗旨，打造优质学术资源的共享数据平台。

版权说明：未经授权严禁任何形式的媒体转载和摘编，并且严禁转载至微信以外的平台！

---

原创文章首发于蔻享学术，仅代表作者观点，不代表蔻享学术立场。

转载请在『蔻享学术』公众号后台留言。

点击阅读原文～发现惊喜！