Title:
MULTI-OBJECTIVE MULTIMODAL PARTICLE SWARM OPTIMIZATION METHOD BASED ON BAYESIAN ADAPTIVE RESONANCE
Document Type and Number:
WIPO Patent Application WO/2022/007376
Kind Code:
A1
Abstract:
Disclosed in the present invention is a multi-objective multimodal particle swarm optimization method based on Bayesian adaptive resonance, comprising: dividing all particles into a plurality of populations using a Bayesian adaptive resonance theory; sorting the particles in the populations according to a non-dominated sorting method and a special crowding distance; updating the particles in the populations using a personal best of the particles and a global best of the populations; connecting non-dominant solution sets of the populations end by end to form a closed ring topology; performing local search using a ring topology-based particle swarm optimization algorithm; repeating the update process and the search process above until a terminal condition is satisfied; and outputting all non-dominant solution sets and Pareto fronts. The present invention is applicable to solve the optimization of multi-objective multimodal problems, and can not only find the distribution of Pareto front in a target space but also find a corresponding Pareto optimal solution set in a decision variable space, thus providing a redundant backup method is provided, and improving the reliability of engineering practice activities.
Inventors:
YANG SHUNKUN (CN)
YAO QI (CN)
YAO QI (CN)
Application Number:
PCT/CN2021/070103
Publication Date:
January 13, 2022
Filing Date:
January 04, 2021
Export Citation:
Assignee:
UNIV BEIHANG (CN)
International Classes:
G06F30/25; G06F30/15; G06F30/27; G06N3/00
Foreign References:
| CN111814251A | 2020-10-23 | |||
| CN106203689A | 2016-12-07 | |||
| CN108268979A | 2018-07-10 | |||
| CN111077896A | 2020-04-28 |
Attorney, Agent or Firm:
MUDA INTELLECTUAL PROPERTY FIRM (CN)
Download PDF: