Research on KPCA-MDT Algorithm for Intelligent Optimization of Coal Mine Drilling Device Fleets

Since the operators of drilling device fleets in underground coal mines are low in efficiency, failing to meet the requirement of downsizing and synergia there, while operational arrangement of rig fleets mainly depends on the experiences of the management, requiring the management has higher expertise, but they are not able to make accurate and reasonable judgment, this paper puts forth a method to intelligently optimize the drilling device fleets by using KPCA-MDT algorithm: first, the characteristics of strata to be operated, and the numbers of workers, drilling devices and drill stems etc. are taken as condition attributes and KPCA algorithm is applied to extract the features of the datasets to eliminate the correlations among all influencing factors; second, the training samples extracted with features are analyzed with MDT to search for the optimal weight of each attribute, and on the basis of local optimization, the classified boundary model of rigs is disturbed stochastically in an effort to find out better boundary; finally, the network optimization model of drilling device fleets aimed to minimize the overall operational period is built, and part of the decision tree is reconstituted to cut the expenditure of training time by using and adjusting the order to divide attributes in the branch path. Through testing with an experimental example at the site, the operational period is shortened by 3.79% and the accuracy of the algorithm rises by 21.2%, achieving great result from optimization. This is a reliable method to intelligently optimize the operation of drilling device fleets in the underground coal mines.