Two-stage and Lagrangian Dual Decision Rules for Multistage Adaptive Robust Optimization
ARSLAN, Ayşe
Formulations étendues et méthodes de décomposition pour des problèmes génériques d'optimisation [EDGE]
Formulations étendues et méthodes de décomposition pour des problèmes génériques d'optimisation [EDGE]
ARSLAN, Ayşe
Formulations étendues et méthodes de décomposition pour des problèmes génériques d'optimisation [EDGE]
< Reduce
Formulations étendues et méthodes de décomposition pour des problèmes génériques d'optimisation [EDGE]
Language
en
Document de travail - Pré-publication
English Abstract
In this work, we design primal and dual bounding methods for multistage adjustable robust optimization (MSARO) problems by adapting two decision rules rooted in the stochastic programming literature. This approach approximates ...Read more >
In this work, we design primal and dual bounding methods for multistage adjustable robust optimization (MSARO) problems by adapting two decision rules rooted in the stochastic programming literature. This approach approximates the primal and dual formulations of an MSARO problem with two-stage models. From the primal perspective, this is achieved by applying two-stage decision rules that restrict the functional forms of a certain subset of decision variables. We present sufficient conditions under which the well-known constraint-and-column generation algorithm can be used to solve the primal approximation with finite convergence guarantees. From the dual side, we introduce a distributionally robust dual problem for MSARO models using their nonanticipative Lagrangian dual and then apply linear decision rules on the Lagrangian multipliers. For this dual approximation, we present a monolithic bilinear program valid for continuous recourse problems, and a cutting-plane method for mixed-integer recourse problems. Our framework is general-purpose and does not require strong assumptions such as a stage-wise independent uncertainty set, and can consider integer recourse variables. Computational experiments on newsvendor, location-transportation, and capital budgeting problems show that our bounds yield considerably smaller optimality gaps compared to the existing methods.Read less <
English Keywords
Optimization under uncertainty
Robust optimization
Decision rules
ANR Project
Bornes primales et duales pour optimisation robuste adjustable - ANR-22-CE48-0018
Origin
Hal imported