基于QPSO的模糊分类系统优化设计

提出了构建模糊分类系统的有效方法.通过量子位选择的方法对初始的模糊规则进行优化,减少种群规模、提高全局搜索能力,且可以大幅缩短训练时间,达到快速收敛、有效分类的目的.为了优化模糊分类空间和减少模糊规则数目,提出了量子行为粒子群优化(QPSO)算法,提高初始模糊分类系统的性能.实验结果证明:优化方法较之其他方法更有效率,准确率更高.

Optimization design of fuzzy classifcation system based on QPSO

An effective method for construction of fuzzy classification system(FCS) is proposed. In FCS,the initial fuzzy rules are optimized with a quantum bit which has many unique advantages such as small population size,fast convergence,short training time and strong global search ability. After then,in order to accomplish the optimization for the fuzzy classification space and reduce number of fuzzy rules,quantum-behaved particle swarm optimization (QPSO) algorithm is proposed to improve characteristics of initial FCS. The experimental result demonstrates that this method is more efficient and accurate than other methods without QPSO.