Title: Time-varying social emotional optimisation algorithm

Authors: Yanchun Liu; Zhendong Xu

Addresses: College of Instrumentation and Electrical Engineering, Jilin University, ChangChun 130061, China; The Aviation University of Air Force, ChangChun 130022, China ' The Aviation University of Air Force, ChangChun 130022, China

Abstract: Social emotional optimisation algorithm (SEOA) is a recently proposed swarm intelligent algorithm by simulating the decision process among human society. In SEOA, each individual denotes one virtual person, and three different kinds of emotions are designed: low-spirited, middle-spirited and high-spirited, then, each person selects the behaviour emotion according to emotional index. In the standard version of SEOA, there are three parameters used to control the influences of personal experiences, social experiences and failure experiences, however, all of them are designed as fixed values. This phenomenon is confused with the nature. In fact, the influences of these experiences are different for different period. For example, individual experiences are more important for the early period, the same as failure experiences, while the social experiences are more important in the later period. Therefore, to meet this phenomenon, a dynamic time-varying strategy is designed. To testify the performance of modified SEOA, three famous benchmarks are chosen, they are Rosenbrock model, Rastrigin model and Griewank model. The dimension is from 30 up to 300. Simulation results show this modification improves the performance significantly especially for multimodal, high-dimensional problems.

Keywords: social emotional optimisation algorithm; SEOA; low-spirited; middle-spirited; high-spirited; time-varying strategy; swarm intelligence; emotions; personal experiences; social experiences; failure experiences.

DOI: 10.1504/IJCSM.2012.051625

International Journal of Computing Science and Mathematics, 2012 Vol.3 No.4, pp.376 - 384

Available online: 23 Jan 2013 *

Full-text access for editors Access for subscribers Purchase this article Comment on this article