Research Of Single Objective And Multi-objective Genetic Algorithm

John Holland, who was a professor in the University of Michigan proposed Genetic algorithm based on the biological evolution theory. Genetic algorithm is an optimization algorithm based on population search. It has been widely used in image processing, system identification, automatic control, economic forecasts and engineering optimization areas due to the algorithm is simple, efficient, and not affected by specific issues. In recent years, the studies of genetic algorithms are mainly concentrated in the execution strategy and the design of genetic operators. The theoretical researches of genetic algorithm mainly are focused on solving low convergence, poor local search capability and improving the accuracy of the optimal solution.Common genetic algorithms may have a tendency to converge towards local optima, and have the characteristics of poor local searching ability and slow global convergence. In order to solve these problems, this paper proposes an improved genetic algorithm. This improved genetic algorithm combines the idea of using evolutionary generation and fitness distribution to adjust the crossover probability and mutation probability. The simulation results demonstrate that the improved genetic algorithm is superior to common genetic algorithms. Compared to common genetic algorithms, the improved genetic algorithm not only has better searching abilities, but can also converge to global optima quickly.In the current rapid development of science and technology, a large number of scientific research and engineering practices belong to multi-objective optimization problems. The method to handle complex multi-objective optimization problems is a hot spot. Among many methods that how to dealing with multi-objective optimization problems by using the multi-objective genetic algorithm (MOGA) has played an important role. Another research work in this paper is to introduce the improvement strategies of single objective genetic algorithm into multi-objective genetic algorithm. The algorithms with elite selection strategy of the dominant ranking (NSGA-II) is a classical multi-objective genetic algorithm. By using NSGA-II as standard multi-objective genetic algorithm, a large number of simulations demonstrate that the performance of the NSGA-II is improved by using the improvement strategies proposed in this paper.Finally, the paper sums up the deficiency of algorithms, raises other improvement approaches and show the development prospects of genetic algorithm.