中文摘要：

为市政的水消费的短期的预报的一条途径基于混乱理论的最大的 Lyapunov 代表被介绍。城市的水消费的时间系列的混乱特征借助于最大的 Lyapunov 代表和关联尺寸被检验。由使用最大的 Lyapunov 代表，为城市的水消费的一个短期的预报模型被开发，它与人工的神经网络(ANN ) 相比是在案例研究的途径。结果显示模型比 ANN 方法，和它的预报吝啬的亲戚基于最大的 Lyapunov 代表有更高的预言精确和预报稳定性当它在规模以外是 60.6% 时，错误在它的最大的可预言的时间规模以内是 9.6% 。

英文摘要：

An approach for short-term forecasting of municipal water consumption was presented based on the largest Lyapunov exponent of chaos theory. The chaotic characteristics of time series of urban water consumption were examined by means of the largest Lyapunov exponent and correlation dimension. By using the largest Lyapunov exponent a short-term forecasting model for urban water consumption was developed, which was compared with the artificial neural network （ANN） approach in a case study. The result indicates that the model based on the largest Lyapunov exponent has higher prediction precision and forecasting stability than the ANN method, and its forecasting mean relative error is 9.6% within its maximum predictable time scale while it is 60.6% beyond the scale.