Volume 51, Number 4, October-December 2017
|Page(s)||1301 - 1315|
|Published online||27 November 2017|
A new hypervolume-based differential evolution algorithm for many-objective optimization∗
1 College of Economics and Management, Beijing University of Technology, Beijing 100124, China.
2 Research Base of Beijing Modern Manufacturing Development, Beijing 100124, China .
3 Institute of Laser Engineering, Beijing University of Technology, Beijing 100124, China .
Received: 17 February 2016
Accepted: 2 March 2017
Evolutionary algorithms are successfully used for many-objective optimization. However, solutions are prone to become nondominated from each other with the increase in the number of objectives, which reduces the efficiency of Pareto dominance-based algorithms. In this paper, a new hypervolume-based differential evolution algorithm (MODEhv) is proposed for many-objective optimization problems (MaOPs). In MODEhv, a modified differential evolution paradigm with automatic parameter configuration strategy is used to balance exploration and exploitation of the algorithm. Besides, the hypervolume indicator is incorporated for the selection of solutions to be varied and solutions to be kept in archive respectively. Finally, a threshold technique is employed to improve diversity of solutions in archive. MODEhv is investigated on a set of widely used benchmark problems and compared with five state-of-the-art algorithms. The experimental results show the efficiency of MODEhv for solving MaOPs.
Mathematics Subject Classification: 65K10 / 68T20
Key words: Differential Evolution / Hypervolume indicator / Many-objective optimization / Many-objective evolutionary algorithm
This paper is supported by the National Natural Science Foundation of China under Grant 61273230, 61603011 and 61603010, Shandong Financial Collaborative Innovation Center under Grant 14AWTJ01-4, Capital Social development and Social Management Collaborative Innovation Center, and Research Base of Beijing Modern Manufacturing Development.
© EDP Sciences, ROADEF, SMAI 2017
Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.
Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.
Initial download of the metrics may take a while.