CAJ | 학술논문

为了提高网络流量预测的精度，提出先进行小波变换后利用LS-SVM的网络流量预测模型，对于LS-SVM参数的优化，提出一种基于模拟退火算法的自适应混沌量子粒子群算法( AS-QPSO)。该算法在量子粒子群算法的基础上加入了自适应和混沌特性，使算法具有动态自适应性，改善算法的全局寻优能力，再引入模拟退火算法避免陷入局部最优，使算法具有更好的收敛性和稳定性。实验结果表明：与其他算法优化的LS-SVM模型相比，该模型具有较好的泛化能力﹑更高的预测精度以及很好的稳定性。
위료제고망락류량예측적정도，제출선진행소파변환후이용LS-SVM적망락류량예측모형，대우LS-SVM삼수적우화，제출일충기우모의퇴화산법적자괄응혼돈양자입자군산법( AS-QPSO)。해산법재양자입자군산법적기출상가입료자괄응화혼돈특성，사산법구유동태자괄응성，개선산법적전국심우능력，재인입모의퇴화산법피면함입국부최우，사산법구유경호적수렴성화은정성。실험결과표명：여기타산법우화적LS-SVM모형상비，해모형구유교호적범화능력﹑경고적예측정도이급흔호적은정성。
For improving the prediction accuracy of network traffic,a new network traffic prediction model is proposed based on wavelet transform and optimized LS-SVM. To optimize the parameters of LS-SVM,a kind of adaptive chaos quantum-behaved particle swarm optimization based on simulated annealing algorithm ( AS-QPSO) is proposed. The algorithm joins adaptive and chaotic characteristics based on QPSO,making it dynamic adaption,improving the capacity of the global optimization. Then the simulated annealing algorithm is introduced to avoid falling into local optimum,the algorithm has better convergence and stability. Experimental results show that com-pared with other algorithm optimized LS-SVM model,the proposed model is more efficient with higher precision,better generalization performance and stability.