When we address an optimization problem whose definition (in some aspect) changes over time, we are in the presence of a Dynamic Optimization Problem (DOP). The aspects that can change over time are the objective function, the domain of the variables, the decision variables or the constraints among others. In this talk, I will provide a brief review of the recent work that my co-authors and I have done on DOPs with cooperative hybrid metaheuristics. On the one hand, I will present a method for continuous DOPs where a set of stochastic local searches cooperates by exchanging information through a central coordinator. And on the other hand, an algorithm portfolio for binary DOPs that incorporates a learning scheme to select the most appropriate solver at each stage of the search.
dynamic problems, continuous optimization, cooperative metaheuristics