Evolutionary multiobjetive optimization in non-stationary environments

This paper proposes an approach, called Multiobjective Algorithm for Dynamic Environments (MADE), which extendes Fonseca and Fleming's MOGA (with an external archive) so that it can deal with dynamic environments. MADE includes two techniques to maintain diversity and also uses specialized functions that implements the dynamism required. In order to validate MADE, we defined a dynamic version of a static test problem (with 3 objectives) previously proposed in the specialized literature. The preliminary results obtained indicate that the proposed approach provides an acceptable response to the type of changes studied.