Accelerating the regularized decomposition method for two stage stochastic linear problems

被引：93

作者：

Ruszczynski, A

Swietanowski, A

机构：

[1] UNIV WISCONSIN,DEPT IND ENGN,MADISON,WI 53706

[2] WARSAW UNIV TECHNOL,INST CONTROL & COMPUTAT ENGN,WARSAW,POLAND

来源：

EUROPEAN JOURNAL OF OPERATIONAL RESEARCH | 1997年 / 101卷 / 02期

关键词：

stochastic programming; decomposition; non-smooth optimization;

D O I：

10.1016/S0377-2217(96)00401-8

中图分类号：

C93 [管理学];

学科分类号：

12 ; 1201 ; 1202 ; 120202 ;

摘要：

Practical improvements of the regularized decomposition algorithm for two stage stochastic problems are presented. They are associated with the primal simplex method for solving subproblems. A penalty formulation of the subproblems is used, which facilitates crash and warm starts, and allows more freedom when creating the model. The computational results are highly encouraging. (C) 1997 Elsevier Science B.V.

引用

页码：328 / 342

页数：15

共 25 条

[1]

[Anonymous], NUMERICAL TECHNIQUES

[2]

BIRGE JR, 1986, MATH PROGRAM STUD, V27, P54, DOI 10.1007/BFb0121114

[3] A MULTICUT ALGORITHM FOR 2-STAGE STOCHASTIC LINEAR-PROGRAMS [J].