Scalable Parallel Benders Decomposition for Stochastic Linear Programming

Abstract We develop a scalable parallel implementation of the classical Benders decomposition algorithm for two-stage stochastic linear programs. Using a primal-dual, path-following algorithm for solving the scenario subproblems we develop a parallel implementation that alleviates the difficulties of load balancing. Furthermore, the dual and primal step calculations can be implemented using a data-parallel programming paradigm. With this approach the code effectively utilizes both the multiple, independent processors and the vector units of the target architecture, the Connection Machine CM-5. The, usually limiting, master program is solved very efficiently using the interior point code LoQo on the front-end workstation. The implementation scales almost perfectly with problem and machine size. Extensive computational testing is reported with several large problems with up to 2 million constraints and 13.8 million variables.

[1]  John R. Birge,et al.  Efficient solution of two-stage stochastic linear programs using interior point methods , 1992, Comput. Optim. Appl..

[2]  K. A. Ariyawansa,et al.  Performance of a benchmark parallel implementation of the Van Slyke and Wets algorithm for two-stage stochastic programs on the Sequent/Balance , 1991, Concurr. Pract. Exp..

[3]  G. Amdhal,et al.  Validity of the single processor approach to achieving large scale computing capabilities , 1967, AFIPS '67 (Spring).

[4]  Y. Censor,et al.  Parallel Optimization: Theory, Algorithms, and Applications , 1997 .

[5]  R. Wets,et al.  Stochastic programming , 1989 .

[6]  John M. Mulvey,et al.  An Extension of the DQA Algorithm to Convex Stochastic Programs , 1994, SIAM J. Optim..

[7]  Stavros A. Zenios,et al.  A Massively Parallel Algorithm for Nonlinear Stochastic Network Problems , 1993, Oper. Res..

[8]  R. Wets,et al.  L-SHAPED LINEAR PROGRAMS WITH APPLICATIONS TO OPTIMAL CONTROL AND STOCHASTIC PROGRAMMING. , 1969 .

[9]  G. Dantzig,et al.  Large-Scale Stochastic Linear Programs: Importance Sampling and Benders Decomposition , 1991 .

[10]  J. Birge,et al.  Computing block-angular Karmarkar projections with applications to stochastic programming , 1988 .

[11]  Stavros A. Zenios,et al.  Solving multistage stochastic network programs on massively parallel computers , 1996, Math. Program..

[12]  Peter Kall,et al.  Stochastic Programming , 1995 .

[13]  R. Wets.  Stochastic Programs with Fixed Recourse: The Equivalent Deterministic Program , 1974 .

[14]  Stavros A. Zenios,et al.  On the Scalability of Data-Parallel Decomposition Algorithms for Stochastic Programs , 1994, J. Parallel Distributed Comput..

[15]  Robert J. Vanderbei,et al.  Symmetric indefinite systems for interior point methods , 1993, Math. Program..

[16]  John M. Mulvey,et al.  A diagonal quadratic approximation method for large scale linear programs , 1992, Oper. Res. Lett..

[17]  Stavros A. Zenios,et al.  Proximal minimizations with D-functions and the massively parallel solution of linear network programs , 1993, Comput. Optim. Appl..

[18]  Elizabeth R. Jessup,et al.  Parallel Factorization of Structured Matrices Arising in Stochastic Programming , 1994, SIAM J. Optim..

[19]  G. B. Dantzig,et al.  Solving stochastic linear programs on a Hypercube multicomputer. Technical report , 1991 .

[20]  J. F. Benders.  Partitioning procedures for solving mixed-variables programming problems , 1962 .