摘要
根据响应面法近似计算在多学科优化中的特点,针对普通神经网络构造响应面方法存在的主要问题,将遗传算法和神经网络结合,设计出混合神经网络并构造了协同优化算法的响应面,在任务调度模块的支持下,可以自主地选择近似和精确计算。采用混合神经网络调度法对玻璃钢拖网渔船进行技术经济论证,结果表明,该方法能够提高优化结果的效率和鲁棒性。
The problem dealing with artificial neural network to construct response surface in multidiscipline optimization,and combination of genetic algorithm with artificial neural network is discussed to design a hybrid artificial neural network to set up response surface for collaborative optimization.The collaborative evaluation may choose automatically approximate and precise computation by the task scheduling model.A practical case of technological and economical collaborative evaluation for FPR trawler vessels demonstrates that the proposed method is of better efficiency and robust.
出处
《大连水产学院学报》
CSCD
北大核心
2010年第2期162-166,共5页
Journal of Dalian Fisheries University
基金
国家渔业船舶检验局项目(国渔检办<2003>44号)
辽宁省教育厅科学技术计划项目(2008146)
关键词
神经网络
遗传算法
响应面
渔船协同论证
artificial neural network
genetic algorithm
response surface
collaborative evaluation of fishing vessels
