摘要
动稳定度是评价沥青混合料在规定条件下抵抗塑性流动变形能力的指标,它的大小直观反映了沥青混合料抗车辙能力的强弱。介绍了BP、遗传算法(GA)优化BP两种神经网络,建立了沥青混合料抗车辙性能预估模型,并以此分别对沥青混合料车辙动稳定度进行预测,实验结果表明:基于遗传算法(GA)优化BP神经网络的沥青混合料动稳定度预测方法,能够使网络收敛速度加快并避免局部极小;GA-BP神经网络在收敛速度和预测精度方面均优于BP神经网络。
Dynamic stability is an index to evaluate the ability of asphalt mixes to resist plastic flow under specified conditions.Its magnitude reflects the anti-rutting ability of asphalt mixes.The paper introduces two neutral network of BP and GA optimal BP,establishes a prediction model of asphalt mixes anti-rutting performance,and makes a prediction of the dynamic stability of asphalt mixes by this model.The results of the tests indicate that: the method of asphalt mixes dynamic stability prediction based on GA optimal BP could speed up the convergence of network and avoid regional minimization;neutral network of GA-BP is better than neutral network BP in term of the speed of convergence and precision of prediction.
出处
《山东建筑大学学报》
2010年第3期226-230,共5页
Journal of Shandong Jianzhu University
基金
山东省教育厅科技计划项目(J07YA07)
山东省自然科学基金项目(ZR2009FM010)
关键词
道路工程
动稳定度
遗传神经网络
预估
road engineering
dynamic stability
genetic algorithm and neural network
prediction