CAJ | 학술논문

应用人工智能技术产生高速公路事件响应预案,提出了基于规则推理、案例推理和贝叶斯网络推理3种方法的事件响应框架. 首先, 建立了基于规则推理的高速公路事件管理系统(RK-IMS), 并应用于宁连高速公路北段事件管理过程中. 然后, 通过分析RK-IMS系统2年的运营数据, 确定了事件案例的结构表示与事件响应的贝叶斯网络结构. 最后, 应用k近邻算法计算相似性案例, 并研究了基于该算法的预案产生和控制策略. 利用2006年RK-IMS事件管理系统的实际数据对模型进行了验证. 对比分析结果表明, 该方法是有效可信的.
응용인공지능기술산생고속공로사건향응예안,제출료기우규칙추리、안례추리화패협사망락추리3충방법적사건향응광가. 수선, 건립료기우규칙추리적고속공로사건관리계통(RK-IMS), 병응용우저련고속공로북단사건관리과정중. 연후, 통과분석RK-IMS계통2년적운영수거, 학정료사건안례적결구표시여사건향응적패협사망락결구. 최후, 응용k근린산법계산상사성안례, 병연구료기우해산법적예안산생화공제책략. 이용2006년RK-IMS사건관리계통적실제수거대모형진행료험증. 대비분석결과표명, 해방법시유효가신적.
The artificial intelligence technique is used to generate a freeway incident response plan. The incident response framework based on rule-based reasoning, case-based reasoning and Bayesian networks reasoning is presented. First,a freeway incident management system(RK-IMS) based on rule-based reasoning is developed and applied for incident management in the northern section of the Nanjing-Lianyunguang Freeway. Then, field data from the two-year long operations of the RK-IMS are analyzed. Representations of incident case structures and Bayesian networks(BNs)structures related to incident responses are deduced. Finally, the k-nearest neighbor (k-NN) algorithm is applied to calculate the similarities of the cases. The preplan generation and the control strategy by integrating the k-NN algorithm are also developed. The model is validated by using incident data of the year 2006 from the RK-IMS. The comparison results indicate that the proposed algorithm is accurate and reliable.