摘要
人工蜜蜂群(ABC)优化算法具有较强的全局搜索能力。在标准算法的基础上,参考粒子群优化算法,加入当前全局最优解对算法的有益引导;当观察蜂在引导蜂所在食物源附近搜索时,引入混沌搜索机制,改善局部搜索性能。利用改进的ABC算法,以网络训练的最小方差F为优化指标,优化神经网络的连接权值。优化后的神经网络用于瓦斯预测,取得了良好的效果。
Artificial bee colony(ABC) optimal algorithm has strong global search ability.Based on standard algorithm,inspired by PSO algorithm,current global optimal solution has been considered.Whenever onlooker bee searches around food source that employed bee searched,chaos search mechanism is introduced,in order to ameliorate local search performance.Using improved ABC algorithm,weights of neural network is optimized,aiming at F of network training as optimal index.Optimized neural network has been used in gas forecast,excellent result has been obtained.
出处
《传感器与微系统》
CSCD
北大核心
2011年第4期79-81,92,共4页
Transducer and Microsystem Technologies
基金
国家自然科学基金资助项目(50874059)
辽宁省重点科技计划资助项目(2007231003)
辽宁省优秀人才基金资助项目(2007R24)
辽宁省创新团队基金资助项目(2007T071)
关键词
人工蜜蜂群优化算法
神经网络
混沌搜索
瓦斯预测
artificial bee colony(ABC) optimal algorithm
neural network
chaos search
gas forecast