The lecture will focus on metaheuristics and will present a general view of these methods used to solve combinatorial optimization problems over the ages. The talk will present a vision of what has been a glorious past (the golden age) of metaheuristics. Over the last 20-30 years, they have been successfully used to solve various optimization problems. Many different techniques such as Simulated Annealing, Tabu Search, Variable Neighborhood Search, GRASP, Genetic Algorithms, Memetic Algorithms and others have shown their efficiency and superiority on many problems. A more recent trend in metaheuristics publications has seen the appearance of too many nature-inspired metaheuristics. All these methods are “polluting” the domain and for most of them do not bring anything from a scientific point of view. The talk will quickly explain why we have reached this situation. The lecture will end with the presentation of recent and future works at the crossroad of metaheuristics and machine learning that should bring a bright future to the metaheuristic community.