G. LEGUIZAMÓN, G. ORDÓÑEZ, and S. MOLINA
Universidad Nacional de San Luis, Argentina
Universidad de Málaga, Spain
Many real-world optimization problems have a nonstationary environment; that is, some of their components depend on time. These types of problems are called dynamic optimization problems (DOPs). Dynamism in real-world problems can be attributed to several factors: some are natural (e.g., weather conditions); others can be related to human behavior (e.g., variation in aptitude of different individuals, inefficiency, absence, and sickness); and others are business related (e.g., the addition of new orders and cancellation of old ones). Techniques that work for static problems may therefore not be effective for DOPs, which require algorithms that make use of old information to find new optima quickly. An important class of techniques for static problems are the metaheuristics (MHs), the most recent developments in approximate search methods for solving complex optimization problems .
In general, MHs guide a subordinate heuristic and provide general frameworks that allow us to create new hybrids by combining different concepts derived from classical heuristics, artificial intelligence, biological evolution, natural systems, and statistical mechanical techniques used to improve ...