X - UWSpace - University of Waterloo
X - UWSpace - University of Waterloo
X - UWSpace - University of Waterloo
Create successful ePaper yourself
Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.
in this algorithm, both subproblems are in equivalent positions and play the role <strong>of</strong> both<br />
the master problem and the subproblem <strong>of</strong> the traditional method At each iteration, the stage one<br />
subproblem gives the upper bound <strong>of</strong> the onginai pmblem by restncting the dual variables to<br />
convex combinations <strong>of</strong> known dual solutions and the stage two subproblem gives a lower bound<br />
by restricting primal vviables to convex combinations <strong>of</strong> known primal solutions. The algorithm<br />
itemtes until the upper bound and Iower bound reach an equilibnum point.<br />
Their tests for eleven problerns with the two stage, block-uiangular stmcnires showed bat<br />
the new method is usually faster and more efficient than the iraditional rnethods.<br />
Park [19%] extended this idea to convex. nonlinear protgmming rnodeis. He proved and<br />
tested that the new algorithm for the two stage case converged to an optimal solution in a finite<br />
number <strong>of</strong> i tentions.<br />
Prirnal-Dual Nested Decornposition Method for Multi-Stage Models<br />
Lm and Fuller [199Sb] also presented a nested primal-dual decomposition aigorithm for<br />
the multi-stage LP problerns with block-trianplar matrix structure. In ihis aipnthrn. the originiil<br />
multi-stage problem is divided into a sequence <strong>of</strong> a pair <strong>of</strong> subproblems for each stage. These<br />
subproblems are coordinated by passing the proposals forward and cuts backwud; the previous<br />
subprobIems pass the proposals to the following subproblerns fonvard in stage numben (which<br />
designates time period in many models) time and the following subproblems provide cuts to the<br />
previous subproblems backwud. This information fiow between subproblerns is shown in Figure<br />
2.1. from h and Fuller [1995b]. The aigorithm can be perceived as the combination <strong>of</strong> the<br />
primd and the dual nested decomposition algorithms. As in the two stage case, the fint<br />
11