Research on Open-pit Mine Vehicle Scheduling Problem with Approximate Dynamic Programming

摘要

Open-pit mine vehicle scheduling problem is mainly about allocating and designing the schedule of trucks and electric forklift under the constraints, which includes the decision of transportation route, distribution of electric forklifts and trucks number at different mining area with the objective of maximizing equipment utilization rate and reducing production cost. In addition, the various real-time practical changes in mining lead open-pit mine vehicle scheduling problem a dynamic scheduling problem. In this paper, the mathematical model of open-pit mines vehicle scheduling problem using continuous time modeling is established. The transport process, characteristics and requirements of vehicle scheduling problem in open-pit mines are analyzed. For large-scale examples, an approximate dynamic programming model is established by ADP algorithm based on Q-Learning. Numerical experiments of different extraction methods of feature vector and update methods of coefficient vector are carried out, and the results are compared with the results by using solver. The experimental results present that the ADP algorithm designed in this paper can effectively solve the large scale open-pit mine vehicle scheduling problem.