Democracy-inspired particle swarm optimizer with the concept of peer groups
Article Type
Research Article
Publication Title
Soft Computing
Abstract
This article proposes to integrate the concept of governance in human society with the nature-inspired particle swarm optimization (PSO) algorithm. A population-based iterative global optimization algorithm, called Democracy-inspired particle swarm optimization with the concept of peer groups (DPG-PSO) has been developed for solving multidimensional, non-linear, non-convex, and multimodal optimization problems by exploiting the concept of the new peer-influenced topology. Here the particles, each of which model a candidate solution of the problem under consideration, are given a choice to follow two possible leaders who have been selected on the basis of a voting mechanism. The leader and the opposition have their influences proportional to the total number of votes polled in their favor. A detailed empirical study comprising tuning of DPG-PSO parameters and its optimizing mechanism has been presented in the paper. The algorithm is tested in a standard benchmark suite consisting of unimodal, multimodal, shifted and rotated functions. DPG-PSO is found to statistically outperform seven other well-known PSO variants in terms of final accuracy and robustness over majority of the test cases, thus, proving itself as an efficient algorithm.
First Page
3267
Last Page
3286
DOI
10.1007/s00500-015-2007-8
Publication Date
6-1-2017
Recommended Citation
Burman, Ritambhar; Chakrabarti, Soumyadeep; and Das, Swagatam, "Democracy-inspired particle swarm optimizer with the concept of peer groups" (2017). Journal Articles. 2565.
https://digitalcommons.isical.ac.in/journal-articles/2565