The Research On Constructing Non-Dominated Set In Multi-Objective Genetic Algorithm

There are many problems of complex system design in engineering, which can be abstract to multi-objective optimization problems. Generally, those objects are conflict. These problems cannot be solved satisfactorily by the traditional measure for which depend on the mathematic characteristic of the objective functions. The genetic algorithm excel at the engineering optimization for its global searching and applicability. Over the last 15 years, there has been an increasing interest in applying genetic algorithm to multi-objective optimization problems, and many work has been achieved in its theoretics and application. Most multi-objective genetic algorithms(MOGA) use Pareto domination to guide the search, so constructing the non-dominated set became important work. The article introduced briefly the history of MOGA and the actual state of MOGA. The paper reduce the run-time and time complexity of the algorithm to constructing the nod-dominated set base on the feature of current MOGA. And according to the experiments, the new measures are better than formers. Three better algorithms to constructing the non-dominated set have been supported in the article, especially in which the runtime complexity of a algorithm is ( M is the number of objects, N is the size of population) . And it is the lowest runtime complexity compared with existed algorithms.