Scenario Tree Generation and Multi-Asset Financial Optimization Problems

Alois Geyer,Michael Hanke,Alex Weissensteiner
DOI: https://doi.org/10.2139/ssrn.1857411
2011-01-01
SSRN Electronic Journal
Abstract:We compare two popular scenario tree generation methods in the context of financial optimization: Moment matching and scenario reduction. Using a simple problem with a known analytic solution, we find that moment matching, when accompanied by a check to ensure the absence of arbitrage opportunities, exactly replicates this solution. On the other hand, even if the scenario trees generated by scenario reduction are arbitrage-free, the solutions to the approximate optimization problem represented by the reduced tree are highly variable. As a conclusion we strongly favor moment matching over scenario reduction for multi-asset financial optimization problems. The published version differs from this working paper version.
What problem does this paper attempt to address?