Evolutionary multi-objective optimization: current state and future challenges

摘要

Summary form only given. There has been an increasing interest in using heuristic search algorithms based on natural selection (the so called "evolutionary algorithms") for solving a wide variety of problems. As in any other discipline, research on evolutionary algorithms has become more specialized over the years, giving rise to a number of subdisciplines. This paper deals with one of the emerging subdisciplines that have become very popular due to its wide applicability: evolutionary multi-objective optimization (EMO). EMO refers to the use of evolutionary algorithms (or even other biologically inspired heuristics) to solve problems with two or more (often conflicting) objectives. Unlike traditional (single objective) problems, multi-objective optimization problems normally have more than one possible solution. Thus, traditional evolutionary algorithms (e.g., genetic algorithms) need to be modified in order to deal with such problems. This talk provides a general overview of this field, including its historical origins, its most significant developments, some of its most important application areas and its current challenges.