×

Partition-based MQHOA for multimodal optimization. (Chinese. English summary) Zbl 1363.68168

Summary: To solve the problem of multimodal optimization, a partition-based multi-scale quantum harmonic oscillator algorithm (MQHOA) is proposed depending on MQHOA’s global optimization characteristic. It divides reasonably a domain into uniform areas, and then Gauss curves with ground state can be constructed according to the lengths of these uniform areas. With the attenuation of standard deviation, the Gauss curves will converge gradually, thus, extreme points can be found quickly. In addition, two strategies comprising fixed wavelength resolution and muti-level resolution are used for practical problems. Experiments are carried out from three aspects including optimization’s accuracy, all extremal points optimization and global multimodal optimization. Compared with the ant colony algorithm, differential evolution algorithms and other mainstream swarm intelligence algorithms, the algorithm has, in addition to its simpleness on setting parameters, superior optimization accuracy, fast convergence and memory property.

MSC:

68T20 Problem solving in the context of artificial intelligence (heuristics, search strategies, etc.)
68Q12 Quantum algorithms and complexity in the theory of computing
90C59 Approximation methods and heuristics in mathematical programming
PDFBibTeX XMLCite
Full Text: DOI