Quantum-behaved particle swarm optimization with generalized space transformation search
- PDF / 718,061 Bytes
- 17 Pages / 595.276 x 790.866 pts Page_size
- 62 Downloads / 196 Views
(0123456789().,-volV)(0123456789(). ,- volV)
METHODOLOGIES AND APPLICATION
Quantum-behaved particle swarm optimization with generalized space transformation search Yiying Zhang1 • Zhigang Jin1
Springer-Verlag GmbH Germany, part of Springer Nature 2020
Abstract Nature-inspired algorithms have been proved to be very powerful methods for complex numerical optimization problems. Quantum-behaved particle swarm optimization (QPSO) is a typical member of nature-inspired algorithms, and it is a simple and effective population-based technique used in numerical optimization. Despite its efficiency and wide use, QPSO suffers from premature convergence and poor balance between exploration and exploitation in solving complex optimization problems. To address these issues, a new evolutionary technique called generalized space transformation search is proposed, and then, we introduce an improved quantum-behaved particle swarm optimization algorithm combined with this new technique in this study. The proposed generalized space transformation search is based on opposition-based learning and generalized opposition-based learning, which can not only improve the exploitation of the current search space but also strengthen the exploration of the neighborhood of the current search space. The improved quantum-behaved particle swarm optimization algorithm employs generalized space transformation search for population initialization and generation jumping. A comprehensive set of 16 well-known unconstrained benchmark functions is employed for experimental verification. The contribution of the generalized space transformation search is empirically verified, and the influence of dimensionality is also investigated. Besides, the improved quantum-behaved particle swarm optimization algorithm is also compared with some typical extensions of QPSO and several competitive meta-heuristic algorithms. Such comparisons suggest that the improved quantum-behaved particle swarm optimization algorithm may lead to finding promising solutions compared to the other algorithms. Keywords Opposition-based learning Particle swarm optimization Space transformation Numerical optimization
1 Introduction Solving a numerical optimization problem is to find an optimal solution from given all of the possible solutions by an optimization method. Optimization methods can be broadly divided into two major categories, i.e., deterministic methods and meta-heuristic methods. Deterministic methods are traditional techniques, which need to select an initial starting point and apply specific mathematical rules
Communicated by V. Loia. & Zhigang Jin [email protected] Yiying Zhang [email protected] 1
School of Electrical and Information Engineering, Tianjin University, Tianjin 300072, China
in search process (Savsani and Savsani 2016). However, deterministic methods are sensitive to initial values (Rao et al. 2011). Thus, deterministic methods can be as an available option for solving simple and ideal optimization problems while they generally fail to solve com
Data Loading...
