基于数据挖掘的民用机场水泥道面维护辅助决策模型

为满足民用机场水泥道面管理系统智能化辅助决策的要求,基于我国26个民用机场水泥道面的356组历史维护决策数据,分析了道面性能属性评价指标间的相关关系,确定了道面状况指数(PCI)、道面等级号(PCN)、板底脱空率和平整度4种道面性能属性评价指标;考虑道面管理者主观需求,提出了可用资金、许用延误、期望效益和工程安全4种管理需求属性,并给出了属性等级及建议划分标准;归纳了8类常用民用机场水泥道面维护措施.在此基础上,采用数据挖掘中的C5.0决策树算法训练了决策树,从而建立了民用机场水泥道面维护辅助决策模型,并开展了评价和应用.评价结果表明,决策模型预测准确性较高;应用案例表明,模型决策结果较为合理,工程应用性较强.

Maintenance Assistant Decision-Making Model of Civil Airport Cement Pavements Based on Data Mining

To meet the requirements of intelligent maintenance assistant decision making of airport cement pavements, 356 sets of valid data from 26 civil airports in China were selected. The correlation of pavement performance indexes was analyzed, and PCI, PCN, void condition and surface roughness were finally confirmed as the pavement performance variables. Considering management requirements of pavements, available funds, allowable delays, expected benefits and project safety were proposed, and their attribute levels were determined respectively. Besides, 8 kinds of common maintenance measures were also summarized. Subsequently, the maintenance decision making tree by using the C5.0 algorithm of the data mining technology was trained to establish the maintenance assistant decision making model. The evaluation and application of the established model were also conducted. The results show that the model is more accurate in forecasting. The results also show that the decision making is reasonable and the engineering application of the model is more feasible.