The Prediction And Dispatching Application Of Trucks State Time In Open-pit Mine Based On Machine Learning

In the open-pit mining and production process,using the transportation equipment efficiently and rationally is an important issue for the development of mining enterprises.As an important parameter to measure road capacity,travel time can be forecasted in real time,which can realize real-time management and control of trucks during operation.It is also a key factor for real-time dispatching and control of open-pit mine trucks.Therefore,by predicting the state of the truck and the running time of the truck and optimizing the truck dispatching model,the utilization rate of truck equipment can be improved,the economic efficiency of the enterprise can be improved,and the intelligent development of the mine can be promoted to realize efficient and intelligent production.The specific work mainly has the following aspects:(1)This paper expounds the machine learning theory and algorithm involved in the truck state recognition and state time prediction in open pit mine.The paper also introduces the theory of truck dispatching optimization,which provides a way to build and solve the dispatching model based on the prediction of truck running time.(2)In view of the various states of the truck in the process of loading and unloading transportation,a state recognition model of the truck in the open pit is established by using the machine learning algorithm.And use the relevant characteristic data collected by the intelligent dispatching control platform to carry out simulation experiments.The results show that Adaboost algorithm has good recognition performance in open pit truck state recognition.(3)According to the running time of trucks,a prediction model of running time of trucks in open-pit mine is established by using machine learning algorithm.The simulation experiment is carried out by using the relevant feature data.For each state running time prediction,the best machine learning model is different.SVM model can be used to predict the running time of no-load and unload States,while RF model can be used to predict the running time of waiting,loading and overload states.(4)The running time of truck is predicted by the prediction model.The prediction results are brought into the dispatching model as parameters,and the dispatching model based on the truck state running time prediction of open pit is established.Based on the actual operation data collected by the intelligent dispatching control platform,the hybrid genetic particle swarm optimization algorithm is used to solve the problem,and a more practical dispatching strategy is obtained.The research provides new methods and solutions for truck dispatching optimization,and lay a foundation for intelligent dispatching,which has important practical value for open-pit mine truck production dispatching.