基于灰色马尔科夫链优化模型的船舶到港量预测

针对船舶到港量总体呈现递增趋势,既受经济等因素影响又存在波动性的特点,采用灰色预测或灰色马尔科夫链预测模型进行船舶到港量预测存在拟合度小和精度低等问题,将三次指数平滑与灰色马尔科夫链预测模型相结合,建立船舶到港量的灰色马尔科夫链优化预测模型。利用宁波舟山港老塘山港区2008—2017年5月份的船舶到港量数据进行预测应用与精度对比,2018年5月船舶到港量预测结果为805艘,残差均值为41.172,相对残差率为7.07%,优化后的灰色马尔科夫链模型残差均值和相对残差率有一定程度的降低,拟合度有一定程度的提高。分析结果表明:优化后的灰色马尔科夫链模型可进一步提高预测的精度,符合实际要求,能为港口锚地规划提供理论支持。

Prediction of Volume of Ship Arrival with Improved Gray Markov Model

The gray prediction or the gray Markov chain prediction model appears to be inaccurate in predicting those volume of ship arrival which shows the generally increasing trend but with considerable volatility following the economic situations. A new ship arrival forecasting model is built by combining Holt-Wnters method with gray Markov forecast model. The model is verified through processing the ship arrival data of May in nearly 10 years from Lao Tangshan Port area in Ningbo-Zhoushan Port, which, for instance, predicts 805 ship arrives in May, 2018 with the residual mean value 41.172 and the relative residual rate 7.07%. The result is superior to that from traditional models, showing a reduction in the mean residual and relative residual rate as well as an increase in fitting.