作者：
Bengio, YoshuaLodi, AndreaProuvost, AntoineEcole Polytech Montreal
Canada Excellence Res Chair Data Sci Decis Making Pavillon Andre Aisenstadt 2920Chemin Tour Montreal PQ H3T 1J4 Canada Mila
Inst Quebecois Intelligence Artificielle Pavillon Andre Aisenstadt 2920Chemin Tour Montreal PQ H3T 1J4 Canada Univ Montreal
Dept Informat & Rech Operat Pavillon Andre Aisenstadt 2920Chemin Tour Montreal PQ H3T 1J4 Canada
This paper surveys the recent attempts, both from the machine learning and operations research communities, at leveraging machine learning to solve combinatorial optimization problems. Given the hard nature of these p...
详细信息
This paper surveys the recent attempts, both from the machine learning and operations research communities, at leveraging machine learning to solve combinatorial optimization problems. Given the hard nature of these problems, state-of-the-art algorithms rely on handcrafted heuristics for making decisions that are otherwise too expensive to compute or mathematically not well defined. Thus, machine learning looks like a natural candidate to make such decisions in a more principled and optimized way. We advocate for pushing further the integration of machine learning and combinatorial optimization and detail a methodology to do so. A main point of the paper is seeing generic optimization problems as data points and inquiring what is the relevant distribution of problems to use for learning on a given task. (C) 2020 Elsevier B.V. All rights reserved.
The Feasibility Pump (fp) is probably the best-known primal heuristic for mixed-integerprogramming. The original work by Fischetti et al. (Math Program 104(1):91-104, 2005), which introduced the heuristic for 0-1 mix...
详细信息
The Feasibility Pump (fp) is probably the best-known primal heuristic for mixed-integerprogramming. The original work by Fischetti et al. (Math Program 104(1):91-104, 2005), which introduced the heuristic for 0-1 mixed-integer linear programs, has been succeeded by more than twenty follow-up publications which improve the performance of the fp and extend it to other problem classes. Year 2015 was the tenth anniversary of the first fp publication. The present paper provides an overview of the diverse Feasibility Pump literature that has been presented over the last decade.
暂无评论