CAJ | 학술논문

交通网络节点和实体不断增长，对交通智能预测和控制提出更高的要求。传统方法采用智能神经网络PID控制方法进行交通控制，在交通并行负载无规则增长下，系统输出不能有效跟踪参考轨线，导致负载调度和控制效果不好。结合智能蜂群仿生算法和粗糙集理论，提出一种基于蜂群引导粗糙集前馈补偿动态博弈论的交通智能预测控制算法，建立一种基于粗糙集理论的前馈补偿动态博弈数学模型，控制参数在动态博弈中实现自适应寻优，迭代修正预测值的不确定性，预测模型收敛到最优解，克服实体无规则增长导致预测控制品质不好的问题。仿真实验表明，采用该控制方法对交通网络进行预测控制，能适用于大规模并行交通网络调度与预测控制，提高控制的鲁棒性。
교통망락절점화실체불단증장，대교통지능예측화공제제출경고적요구。전통방법채용지능신경망락PID공제방법진행교통공제，재교통병행부재무규칙증장하，계통수출불능유효근종삼고궤선，도치부재조도화공제효과불호。결합지능봉군방생산법화조조집이론，제출일충기우봉군인도조조집전궤보상동태박혁론적교통지능예측공제산법，건립일충기우조조집이론적전궤보상동태박혁수학모형，공제삼수재동태박혁중실현자괄응심우，질대수정예측치적불학정성，예측모형수렴도최우해，극복실체무규칙증장도치예측공제품질불호적문제。방진실험표명，채용해공제방법대교통망락진행예측공제，능괄용우대규모병행교통망락조도여예측공제，제고공제적로봉성。
Traffic network node and entity is growing, the traffic intelligent prediction and control needs higher requirement. The traditional method uses intelligent neural network PID control method of traffic control, traffic load in parallel irregular growth, the output of the system cannot effectively track a reference trajectory, causes the load scheduling and control effect is not good. Combined with the theory of intelligent swarm bionic algorithm and rough set, we proposes a rough set based on bee colony guide traffic intelligent feed forward compensation dynamic game theory of predictive control algorithm, the es-tablishment of a feed forward compensation dynamic game model based on rough set theory. Forecasting model converge to the optimal solution is obtained, and it overcomes the entity irregular growth leading to predictive control not good quality problems. Simulation results show that, using the control method of predictive control of traffic network, it can be suitable for large-scale parallel traffic network scheduling and control, to improve the robustness of the control.