摘要：Most industrial planning and scheduling problems are NP-hard, stochastic, and subject to multi-objective. A wide variety of heuristic methods have been designed or adapted to solve them. However, the Genetic Algorithms (GA) family is both the most used and one of the most efficient for several well-known problems. This paper reviews GAs proposed in the literature, focusing on the techniques to overcome scheduling challenges (cycle avoidance and feasibility). This paper also has a didactic purpose and details modern approaches to reach high-quality solutions: self-adaptation, learning process, diversity-maintenance, parallel computation, multi-objective, and hybridization. These mechanisms are also essential to integrate the method in current IT systems.
