A. J. NEBRO, J. J. DURILLO, F. LUNA, and E. ALBA
Universidad de Málaga, Spain
Many sectors of industry (e.g., mechanical, chemistry, telecommunication, environment, transport) are concerned with large and complex problems that must be optimized. Such problems seldom have a single objective; on the contrary, they frequently have several contradictory criteria or objectives that must be satisfied simultaneously. Multiobjective optimization is a discipline focused on the resolution of these types of problems.
As in single-objective optimization, the techniques to solve a multiobjective optimization problem (MOP) can be classified into exact and approximate (also named heuristic) algorithms. Exact methods such as branch and bound [18,22], the A* algorithm , and dynamic programming  are effective for problems of small size. When problems become more difficult, usually because of their NP-hard complexity, approximate algorithms are mandatory.
In recent years an approximate optimization technique known as metaheuristics has become an active research area [1,9]. Although there is not a commonly accepted definition of metaheuristics , they can be considered high-level strategies that guide a set of simpler techniques in the search for an optimum. Among these techniques, evolutionary algorithms for ...