Summary
International Symposium on Nonlinear Theory and its Applications
2017
Session Number:B2L-C
Session:
Number:B2L-C-5
Study of Si Algorithm That Individual to Dropout
Takuya Shindo, Kenya Jin'no,
pp.510-513
Publication Date:2017/12/4
Online ISSN:2188-5079
Summary:
A swarm of creatures such as birds, fish and ants may behave like having intelligence. The intelligence found in the behavior of such swarm is called swarm intelligence (SI). Many of the optimization algorithms such as Particle Swarm Optimization (PSO) and Ant Colony Optimization (ACO) using SI are inspired by the behavior of actual swarm of organisms. In the framework of optimization, SI is thought to be closely related to evolutionary computation, neural networks, and the like. Therefore, we consider using dropout as one of the neural net- work methods for individuals in the swarm of PSO. PSO has a network structure to exchange information with each other among individuals. The optimal solution search performance change by this network topology has been investigated and reported. The network structure is also considered to be an important factor in dropout. Therefore, in this paper, we investigate the performance change due to the network structure of PSO in using dropout. In addition, we confirmed the performance of the proposed method by numerical simulation.