Comparison Study Of Multiobjective Evolutionary Algorithms

Abstract

Many real-world problems involve two types of difficulties: 1) multiple, conflicting objectives and 2) a highly complex search space. Efficient evolutionary strategies have been developed to deal with both difficulties. Evolutionary algorithms possess several characteristics such as parallelism and robustness that make them preferable to classical optimization methods. In the presented paper I conducted comparison studies among the well-known evolutionary algorithms based on NP-hard 0-1 multiobjective knapsack problem. The great importance of elitism strategy in the multiobjective evolutionary search was confirmed experimentally.