Asynchronous Lagrangian scenario decomposition

被引:8
作者
Aravena, Ignacio [1 ]
Papavasiliou, Anthony [2 ]
机构
[1] Lawrence Livermore Natl Lab, Livermore, CA 94550 USA
[2] Catholic Univ Louvain, CORE, Louvain La Neuve, Belgium
关键词
Asynchronous algorithm; Stochastic programming; Unit commitment; High performance computing; UNIT COMMITMENT; DUAL DECOMPOSITION; OPTIMIZATION; ALGORITHM; GENERATION; MARKETS; POWER;
D O I
10.1007/s12532-020-00185-4
中图分类号
TP31 [计算机软件];
学科分类号
081202 ; 0835 ;
摘要
We present a distributed asynchronous algorithm for solving two-stage stochastic mixed-integer programs (SMIP) using scenario decomposition, aimed at industrial-scale instances of the stochastic unit commitment (SUC) problem. The algorithm is motivated by large differences in run times observed among scenario subproblems of SUC instances, which can result in inefficient use of distributed computing resources by synchronous parallel algorithms. Our algorithm performs dual iterations asynchronously using a block-coordinate subgradient descent method which allows performing block-coordinate updates using delayed information, while candidate primal solutions are recovered from the solutions of scenario subproblems using heuristics. We present a high performance computing implementation of the asynchronous algorithm, detailing the operations performed by each parallel process and the communication mechanisms among them. We conduct numerical experiments using SUC instances of the Western Electricity Coordinating Council system with up to 1000 scenarios and of the Central Western European system with up to 120 scenarios. We also conduct numerical experiments on generic SMIP instances from the SIPLIB library (DCAP and SSLP). The results demonstrate the general applicability of the proposed algorithm and its ability to solve industrial-scale SUC instances within operationally acceptable time frames. Moreover, we find that an equivalent synchronous parallel algorithm would leave cores idle up to 80.4% of the time on our realistic test instances, an observation which underscores the need for designing asynchronous optimization schemes in order to fully exploit distributed computing on real world applications.
引用
收藏
页码:1 / 50
页数:50
相关论文
共 53 条
  • [1] A finite branch-and-bound algorithm for two-stage stochastic integer programs
    Ahmed, S
    Tawarmalani, M
    Sahinidis, NV
    [J]. MATHEMATICAL PROGRAMMING, 2004, 100 (02) : 355 - 377
  • [2] Dynamic capacity acquisition and assignment under uncertainty
    Ahmed, S
    Garcia, R
    [J]. ANNALS OF OPERATIONS RESEARCH, 2003, 124 (1-4) : 267 - 283
  • [3] Ahmed S., 2015, SIPLIB STOCHASTIC IN
  • [4] A scenario decomposition algorithm for 0-1 stochastic programs (vol 41, pg 565, 2013)
    Ahmed, Shabbir
    [J]. OPERATIONS RESEARCH LETTERS, 2015, 43 (02) : 215 - 217
  • [5] A scenario decomposition algorithm for 0-1 stochastic programs
    Ahmed, Shabbir
    [J]. OPERATIONS RESEARCH LETTERS, 2013, 41 (06) : 565 - 569
  • [6] Improving the Integer L-Shaped Method
    Angulo, Gustavo
    Ahmed, Shabbir
    Dey, Santanu S.
    [J]. INFORMS JOURNAL ON COMPUTING, 2016, 28 (03) : 483 - 499
  • [7] [Anonymous], 2001, Stud. Appl. Math.
  • [8] [Anonymous], 2015, Advances in Neural Information Processing Systems
  • [9] Two "well-known" properties of subgradient optimization
    Anstreicher, Kurt M.
    Wolsey, Laurence A.
    [J]. MATHEMATICAL PROGRAMMING, 2009, 120 (01) : 213 - 220
  • [10] ARAVENA I, 2017, 2017 IEEE MANCH POW, pNI398