区间客运需求组合预测模型研究

Research on Interval Combination Forecasting Model of Passenger Demand

考虑客运需求受不确定因素影响,导致基于点预测的值不能客观反映未来需求波动。文章采用区间集合方法表示需求不确定,同时结合组合预测方法,建立实际观察值与预测区间中心值偏差最小、预测区间宽度最小的多目标区间需求组合预测模型。以2003—2013年扬中市客运数据为基础进行预测,实例分析表明区间预测结果不仅能包含实际观察值,还能体现未来需求的不确定性,同时误差平方和均值和平均相对误差指标下降也表明区间组合预测模型精度提高,说明该模型的可行性和有效性。

Passenger demand is influenced by uncertain factors, resulting in disability to reflect future fluctuations objectively through forecasting on the basis of point values. It is critical to handle programs of road construction considering decision risks on the basis of demand uncertainty. Demand uncertainty is expressed as interval sets and the forecast model of interval combination is set up with multi-objective. The passenger demand data of Yangzhong city among 2003-2013 is used in the model. The result shows that the model not only includes the actual values, but also reflects the uncertainty. The paper also examines the results with SSE and MAPE index and good precision in forecasting passenger demand is indicated. The model is proved feasible and effective.