EURO-Online login
- New to EURO? Create an account
- I forgot my username and/or my password.
- Help with cookies
(important for IE8 users)
269. A hybrid decision-making support method for large-scale multi-objective mixed integer programming problems
Invited abstract in session TC-45: Integer Programming for Decision Support, stream Decision Support Systems.
Tuesday, 12:30-14:00Room: 30 (building: 324)
Authors (first author is the speaker)
| 1. | Janusz Miroforidis
|
| Department of Decision Support in the Presence of Risk, Systems Research Institute, Polish Academy of Sciences |
Abstract
Many practical multi-criteria decision-making problems are formulated using Multi-Objective Mixed Integer Programming (MOMIP). Among many methods of multi-criteria decision-making, interactive ones are considered the most rational. In interactive methods, Chebyshev scalarization and MIP solvers can be used to derive efficient outcomes. However, for a large-scale instance of the MOMIP problem, even commercial solvers may not provide the optimal solution of its scalarization within a time limit, and its existence is justified in practice. In such a case, the applicability of interactive methods becomes questionable because this interactivity is lost when the time limit is on the order of many minutes. In this work, for large-scale instances of the MOMIP problem, a hybrid decision-making support method is proposed. It combines features of interactive and a posteriori methods. Before the decision-making process begins, a finite two-sided Pareto front approximation is derived with the help of a MIP solver. It is used to calculate the so-called interval representations of Pareto optimal outcomes designated, directly or not, by weights of Chebyshev scalarization. Decision-makers can use them to reveal their preferences without additional optimization calculations. The use of the proposed method is demonstrated with selected large-scale instances of the MOMIP problem with two and three objective functions. Its limitations and possible enhancements are also discussed.
Keywords
- Programming, Multi-Objective
- Programming, Mixed-Integer
- Multi-Objective Decision Making
Status: accepted
Back to the list of papers