Combined prediction model of truck multi-section travel time in open-pit mine based on velocity field
-
摘要: 受限于露天矿道路的复杂性，现有的卡车行程时间预测方法在实际部署中存在困难，导致卡车优化调度系统只实现调度而非优化。提出了一种基于速度场的露天矿卡车多路段行程时间组合预测模型。将露天矿道路划分为多个路段，采用随机森林算法构建单元预测模型，预测卡车在每一路段的行驶时间，再对各单元预测模型预测值累加，得出卡车在复合路段上的行程时间预测值。为提高预测精度，将卡车平均速度作为行程时间影响因素，根据已采集的卡车速度信息构建速度场，求取路段上所有点卡车速度的平均值，将其近似为卡车在该路段的平均速度并输入单元预测模型。以伊敏露天矿卡车调度系统中的卡车行程信息为基础数据，训练得到组合预测模型，并对该模型进行预测精度与实时性实验，结果表明：基于速度场的露天矿卡车多路段行程时间组合预测模型对于复合路段上的卡车行程时间具有较高的预测精度，平均绝对误差百分比为4.81%，较基于随机森林算法的单一预测模型降低2%以上；组合预测模型运算时间不超过1 s，可实现卡车行程时间实时预测。Abstract: Due to the complexity of road in open-pit mine, the existing truck travel time prediction methods are difficult in the actual deployment. This leads to the truck optimal scheduling system only realizing scheduling instead of optimization. A combined prediction model of truck multi-section travel time in open pit mine based on velocity field is proposed. The open-pit mine road is divided into multiple road sections. the random forest algorithm is used to construct the unit prediction model to predict the travel time of the truck in each section. Then the predicted values of the unit prediction models are accumulated to obtain the travel time predicted value of the truck on in the composite road section. In order to improve the prediction precision, the average velocity of the truck is taken as an influence factor of the travel time. The velocity field is constructed according to the collected velocity information of the truck. The average value of the truck velocity at all points on a road section is calculated, which is approximate to the average velocity of the truck on the road section, and the average velocity is input into the unit prediction model. Based on the data of truck schedule information in truck dispatching system of Yimin Open-pit Mine, the combined prediction model is trained, and the prediction precision and real-time performance of the model are tested. The results show that the combined prediction model of truck travel time in multiple sections of open-pit mine based on velocity field has high prediction precision for truck travel time in composite road sections. The average absolute error percentage is 4.81%, which is more than 2% lower than the single prediction model based on random forest algorithm. The operation time of the combined prediction model is less than 1 s, which can realize the real-time prediction of truck travel time.
-
图 1 露天矿卡车多路段行程时间组合预测模型原理
Figure 1. Principle of combined prediction model of multi section travel time of truck in open-pit mine
图 4 组合预测模型的预测误差分布
Figure 4. Prediction error distribution of the combined prediction model
-
[1] 白润才. 露天矿卡车实时调度优化决策系统及应用效果预测研究[D]. 阜新: 辽宁工程技术大学, 2005.BAI Runcai. Research on truck real-time dispatch optimizing decision system and forecast of applied effect in surface mine[D]. Fuxin: Liaoning Technical University, 2005. [2] SHBAKLO S A, KRUEGER L B. A pilot application study of corridor performance indicators[C]//The Proceedings of the Sixth National Conference on Transportation Planning for Small and Medium-Sized Communities, Washington, 1998: 16-18. [3] COWLING P,JOHANSSON M. Using real time information for effective dynamic scheduling[J]. European Journal of Operational Research,2002,139(2):230-244. doi: 10.1016/S0377-2217(01)00355-1 [4] 孟小前. 基于行程时间实时预测的露天矿调度服务研究[D]. 北京: 中国矿业大学（北京）, 2014.MENG Xiaoqian. Research on scheduling services of open-pit mine based on real-time travel time prediction[D]. Beijing: China University of Mining and Technology（Beijing）, 2014. [5] 唐少虎,刘小明,陈兆盟,等. 基于计算实验的城市道路行程时间预测与建模[J]. 自动化学报,2015,41(8):1516-1527. doi: 10.16383/j.aas.2015.c140846TANG Shaohu,LIU Xiaoming,CHEN Zhaomeng,et al. Urban road travel time prediction and modeling via computational experiments[J]. Acta Automatica Sinica,2015,41(8):1516-1527. doi: 10.16383/j.aas.2015.c140846 [6] VANAJAKSHI L, RILETT L R. A comparison of the performance of artificial neural networks and support vector machines for the prediction of traffic speed[C]// Proceedings of the Intelligent Vehicles Symposium, Parma, 2004: 194-199. [7] MEECH J,PARREIRA J. An interactive simulation model of human drivers to study autonomous haulage trucks[J]. Complex Adaptive Systems,2011,6(11):118-123. [8] OZDEMIR B,KUMRAL M. Appraising production targets through agent-based Petri net simulation of material handling systems in open-pit mines[J]. Simulation Modelling Practice and Theory,2018,87(3):138-154. [9] CHANDA E K,GARDINER S. A comparative study of truck cycle time prediction methods in open-pit mining[J]. Engineering,Construction and Architectural Management,2010,17(5):446-460. doi: 10.1108/09699981011074556 [10] SUN Xiaoyu,ZHANG Hang. The use of a machine learning method to predict the real-time link travel time of open-pit trucks[J]. Mathematical Problems in Engineering,2018,4(4):1-14. [11] ERARSLAN K. Modelling performance and retarder chart of highway trucks by cubic splines for cycle time estimation[J]. Mining Technology,2013,114:161-166. [12] 史文中. 空间数据与空间分析不确定性原理[M]. 北京: 科学出版社, 2015.SHI Wenzhong. Uncertainty principle of spatial data and spatial analysis[M]. Beijing: Science Press, 2015. [13] SILVESTER S A,LOWNDES I S,HARGREAVES D M. A computational study of particulate emissions from an open-pit quarry under neutral atmospheric conditions[J]. Atmospheric Environment,2009,43(40):6415-6424. doi: 10.1016/j.atmosenv.2009.07.006 [14] TSAPAKIS I,CHENG T,BOLBOL A. Impact of weather conditions on macroscopic urban travel times[J]. Journal of Transport Geography,2013,28:204-211. doi: 10.1016/j.jtrangeo.2012.11.003 [15] DINDARLOO S R,SIAMI-IRDEMOOSA E. Determinants of fuel consumption in mining trucks[J]. Energy,2016,112(1):232-240. [16] MOHRI M, ROSTAMIZADEH A. Foundations of machine learning[M]. Cambridge: MIT Press, 2012. [17] KIM S Y,UPNEJA A. Predicting restaurant financial distress using decision tree and AdaBoosted decision tree models[J]. Economic Modelling,2014,36(1):354-362. [18] BREIMAN L I,FRIEDMAN J H. Classification and regression trees[J]. Encyclopedia of Ecology,2015,57(3):582-588. [19] SPOKOINY V,WILLRICH N. Bootstrap tuning in Gaussian ordered model selection[J]. The Annals of Statistics,2019,47(3):1351-1380. -
下载:
点击查看大图
图(6)
计量
- 文章访问数: 178
- HTML全文浏览量: 26
- PDF下载量: 25
- 被引次数: 0
