Acessibilidade / Reportar erro

BIOBJECTIVE INTEGER STOCHASTIC OPTIMIZATION OVER THE INTEGER STOCHASTIC EFFICIENT SET

ABSTRACT

In various Multi-objective Programming (MOP) problems, decision-makers are often faced with a large set of efficient solutions, presenting a challenge in discerning and selecting the best solutions within this broad set. To overcome this, decision-makers can tune and optimize their preference function over this efficient set to identify the optimal solutions that align with their preferences. Most existing methods address such problems in the deterministic case, while we aim to tackle a new challenge by dealing with these problems in a stochastic environment, which increases their complexity. Indeed, we focus in this study on proposing an exact method to optimize two preference functions over the efficient set of a Multi-objective Stochastic Integer Linear Programming (MOSILP) problem in order to find the best compromise solutions without enumerating all elements of this efficient set. The proposed method is based on the combination of two techniques: the first one is called the L-shaped method, while the second one is an adaptation of the branch-and-bound strategy by reinforcing it with efficient cuts and tests, which allows the method to remove a large number of inefficient solutions within the search tree. Moreover, the method exhibits adaptability in addressing the general case involving multiple preference functions. A didactic example is presented for illustration, followed by a computational study evaluating the method’s efficacy and performance by solving randomly generated instances.

Keywords:
multi-objective programming; bi-objective programming; stochastic programming; optimization over an efficient set; two-stage stochastic programming; branch-and-cut

Sociedade Brasileira de Pesquisa Operacional Rua Mayrink Veiga, 32 - sala 601 - Centro, 20090-050 Rio de Janeiro RJ - Brasil, Tel.: +55 21 2263-0499, Fax: +55 21 2263-0501 - Rio de Janeiro - RJ - Brazil
E-mail: sobrapo@sobrapo.org.br