A numerical model has been established to calculate the primary spacing of cellular or dendritic structure with fluid flow considered. The computing results show that the primary spacing depends on the growing velocit...A numerical model has been established to calculate the primary spacing of cellular or dendritic structure with fluid flow considered. The computing results show that the primary spacing depends on the growing velocity, the temperature gradient on the interface and fluid flow. There is a critical growing velocity for the cell-dendrite transition, which has a relationship with the temperature gradient: Rcr = (3-4)×10-9GT. Fluid flow leads to an increase of the primary spacing for dendritic growth but a decrease for cellular growth, resulting in an instability on the interface.展开更多
汇率预测非常困难,其波动具有时变性、随机性和模糊性等统计特征.现存文献中各种方法和模型的预测效果受很多因素影响,其预测力都不及随机游走模型,这就是汇率预测领域所谓的"米斯和罗格夫之谜(The Meese and Rogoff puzzle)"...汇率预测非常困难,其波动具有时变性、随机性和模糊性等统计特征.现存文献中各种方法和模型的预测效果受很多因素影响,其预测力都不及随机游走模型,这就是汇率预测领域所谓的"米斯和罗格夫之谜(The Meese and Rogoff puzzle)".本文使用非参数方法研究汇率波动及其预测模型,发现较之任何参数方法、半参数方法都具有更大的灵活性.为了克服"维数魔咒",本文提出非参数可加模型来研究汇率预测问题.与现有模型相比,在同样的观察样本期内,非参数可加汇率预测模型有更好的样本外预测能力,这有力地证明了"米斯和罗格夫之谜"并非难以破解.此外,我们将非参数可加汇率模型应用于人民币对美元的汇率预测,其结果仍然揭示了该模型很好的拟合度和预测能力.本文为汇率预测这一研究领域提供了新的研究思路和方法.展开更多
文摘A numerical model has been established to calculate the primary spacing of cellular or dendritic structure with fluid flow considered. The computing results show that the primary spacing depends on the growing velocity, the temperature gradient on the interface and fluid flow. There is a critical growing velocity for the cell-dendrite transition, which has a relationship with the temperature gradient: Rcr = (3-4)×10-9GT. Fluid flow leads to an increase of the primary spacing for dendritic growth but a decrease for cellular growth, resulting in an instability on the interface.
文摘汇率预测非常困难,其波动具有时变性、随机性和模糊性等统计特征.现存文献中各种方法和模型的预测效果受很多因素影响,其预测力都不及随机游走模型,这就是汇率预测领域所谓的"米斯和罗格夫之谜(The Meese and Rogoff puzzle)".本文使用非参数方法研究汇率波动及其预测模型,发现较之任何参数方法、半参数方法都具有更大的灵活性.为了克服"维数魔咒",本文提出非参数可加模型来研究汇率预测问题.与现有模型相比,在同样的观察样本期内,非参数可加汇率预测模型有更好的样本外预测能力,这有力地证明了"米斯和罗格夫之谜"并非难以破解.此外,我们将非参数可加汇率模型应用于人民币对美元的汇率预测,其结果仍然揭示了该模型很好的拟合度和预测能力.本文为汇率预测这一研究领域提供了新的研究思路和方法.