Date: November, 17th, 16h00
Place: Inria Lille-Nord Europe
Abstract: In optimization problems, it is common the case where the value of some variables is not well known, or it is uncertain. Traditionally, this uncertainty has been dealt with, from the point of view of Stochastic Optimization, which models these variables by means of probability distributions. However, in the last decade fuzzy sets, and more specifically fuzzy numbers, have increased their popularity as an alternative to Stochastic Optimization in scenarios where not much data is available and probability distributions cannot be inferred. This has been the case in the literature of scheduling problems, where many papers on “fuzzy scheduling” have been appearing since the year 2000. This seminar provides a brief introduction to the use of fuzzy numbers in scheduling problems, and how it can affect the design of some common-use meta-heuristics such as a local search or a genetic algorithm.