In view of the disappointing phenomenon that genetic algorithm is trapped into the local minimum in application of complex problems easily, the double thresholds are introduced to dynamically adjust the similarity of the parents and mutation probability. The proposed algorithm helps to enhance the crossover effectiveness and the population diversity, improving the search efficiency of the algorithm. Besides, the added optimal protection guarantees the optimal individual undestroyed while expanding the population searching area. After all, the improved genetic algorithm is tested by using 164-point TSP model. The experimental results show that the improved genetic algorithm find new resolution and improve the searching efficiency when the population evolution stagnates. And comparative simulations with parameter pairs could provide the theoretical instructions of selecting the thresholds and coefficients for scholars.

Date Published: 27 November 2019
PDF: 7 pages
Proc. SPIE 11321, 2019 International Conference on Image and Video Processing, and Artificial Intelligence, 113212Z (27 November 2019); doi: 10.1117/12.2547974

Published in SPIE Proceedings Vol. 11321:
2019 International Conference on Image and Video Processing, and Artificial Intelligence
Ruidan Su, Editor(s)