A deep learning approach for solving linear programming problems

被引:8
|
作者
Wu, Dawen [1 ]
Lisser, Abdel [1 ]
机构
[1] Univ Paris Saclay, CNRS, CentraleSupelec, Lab Signaux & Syst, F-91190 Gif sur yvette, France
关键词
Linear programming; ODE systems; Neural networks; Deep learning; NEURAL-NETWORK; MACHINE;
D O I
10.1016/j.neucom.2022.11.053
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
Finding the optimal solution to a linear programming (LP) problem is a long-standing computational problem in Operations Research. This paper proposes a deep learning approach in the form of feed -forward neural networks to solve the LP problem. The latter is first modeled by an ordinary differential equations (ODE) system, the state solution of which globally converges to the optimal solution of the LP problem. A neural network model is constructed as an approximate state solution to the ODE system, such that the neural network model contains the prediction of the LP problem. Furthermore, we extend the capability of the neural network by taking the parameter of LP problems as an input variable so that one neural network can solve multiple LP instances in a one-shot manner. Finally, we validate the pro-posed method through a collection of specific LP examples and show concretely how the proposed method solves the example.(c) 2022 Elsevier B.V. All rights reserved.
引用
收藏
页码:15 / 24
页数:10
相关论文
共 50 条
  • [1] A geometric approach for solving fuzzy linear programming problems
    M. R. Safi
    H. R. Maleki
    E. Zaeimazad
    Fuzzy Optimization and Decision Making, 2007, 6 : 315 - 336
  • [2] A geometric approach for solving fuzzy linear programming problems
    Safi, M. R.
    Maleki, H. R.
    Zaeimazad, E.
    FUZZY OPTIMIZATION AND DECISION MAKING, 2007, 6 (04) : 315 - 336
  • [3] Deep learning methods for solving linear inverse problems: Research directions and paradigms
    Bai, Yanna
    Chen, Wei
    Chen, Jie
    Guo, Weisi
    SIGNAL PROCESSING, 2020, 177 (177)
  • [4] A new neural network for solving linear programming problems
    Cichocki, A
    Unbehauen, R
    Weinzierl, K
    Holzel, R
    EUROPEAN JOURNAL OF OPERATIONAL RESEARCH, 1996, 93 (02) : 244 - 256
  • [5] IMPLEMENTATION OF AN INEXACT APPROACH TO SOLVING LINEAR SEMIINFINITE PROGRAMMING-PROBLEMS
    LIN, CJ
    YANG, EK
    FANG, SC
    WU, SY
    JOURNAL OF COMPUTATIONAL AND APPLIED MATHEMATICS, 1995, 61 (01) : 87 - 103
  • [6] Solving Sparse Linear Inverse Problems in Communication Systems: A Deep Learning Approach With Adaptive Depth
    Chen, Wei
    Zhang, Bowen
    Jin, Shi
    Ai, Bo
    Zhong, Zhangdui
    IEEE JOURNAL ON SELECTED AREAS IN COMMUNICATIONS, 2021, 39 (01) : 4 - 17
  • [7] An analysis of a class of neural networks for solving linear programming problems
    Chong, EKP
    Hui, S
    Zak, SH
    IEEE TRANSACTIONS ON AUTOMATIC CONTROL, 1999, 44 (11) : 1995 - 2006
  • [8] A Study on Deep Learning Approach to Optimize Solving Construction Problems
    Roshon, Phillip
    Yang, Feng-Jen
    2021 INTERNATIONAL CONFERENCE ON COMPUTATIONAL SCIENCE AND COMPUTATIONAL INTELLIGENCE (CSCI 2021), 2021, : 170 - 174
  • [9] A new algorithm for solving linear programming problems
    Ramirez, A. L.
    Buitrago, O.
    Britto, R. A.
    Fedossova, A.
    INGENIERIA E INVESTIGACION, 2012, 32 (02): : 68 - 73
  • [10] A NEW APPROACH FOR SOLVING FULLY FUZZY LINEAR FRACTIONAL PROGRAMMING PROBLEMS USING THE MULTI-OBJECTIVE LINEAR PROGRAMMING
    Das, Sapan Kumar
    Mandal, Tarni
    Edalatpanah, S. A.
    RAIRO-OPERATIONS RESEARCH, 2017, 51 (01) : 285 - 297