Research On The Traffic Pattern Recognition And Scheduling Control Of Group Control Elevators

With the growing number of high-rise buildings, the elevator, as an essential perpendicular conveyance of modern buildings, is getting more and more attention. People's requirements on the performance and service quality of the elevator are also increasing. A qualified elevator system, in the process of running, should not only consider the passengers'needs of going to their objective floor, but also the psychological changes of the passengers during their waiting and boarding period, as well as the problem of system energy consumption and others. The development of computer technology has provided a hardware foundation for the application of intelligent algorithm in elevator technologies, however, there is still no effective way to reasonably arrange and schedule elevator resources. Studies on these areas of the elevator system are still in its infancy and there is considerable research space. Therefore, algorithm studies on elevator scheduling have important social and economic significance.First, this paper studies the status and development situation of the group control elevator in the elevator industry. Then from the perspective of group control elevator system, through making mathematical models of the group control elevator system, this paper summarizes the common traffic pattern of the group control elevator system. And then, applying the Random Forest algorithm into the elevator group control system can accurately identify the traffic pattern of the group control elevator system. According to different traffic patterns, it adjusts the evaluation index parameters and forms the comprehensive evaluation index. Then, this paper makes simulation experiments on group control elevator dispatching scheme based on the Hungarian Algorithm to optimize the comprehensive evaluation index. The simulation results show that the Random Forest-Hungarian Algorithm is effective in the traffic pattern recognition and scheduling scheme determination of group control elevator system and it can reduce the waiting time, shorten the running time, reduce system energy consumption, improve the group control elevator system's overall performance and make the group control elevator system to be more intelligent and humane.This paper includes the following aspects:First, through reading a large number of domestic and foreign related materials, this paper studies the development trends and problems urgently need to be addressed of the group control elevator system from an international perspective based on analyzing the working principle of the group control elevator system. Through various methods of analysis and comparison, it ultimately adopts the group control elevator scheduling scheme that combines Random Forests Algorithm with Hungarian Algorithm.Second, taking the actual elevator for example, through analysis on the passenger flow distribution during the operation, the traffic patterns of the group control elevator system are divided into four categories:up-peak traffic pattern, down-peak traffic pattern, the interfloor traffic pattern and idle traffic pattern. After analysis of the characteristics of these four traffic patterns, this paper uses a pattern recognition algorithm to recognize the traffic pattern of the group control elevator, and then uses a scheduling algorithm to determine the scheduling scheme of the group control elevator.Third, this paper makes a more in-depth analysis and study of the group control elevator system. According to different traffic patterns, considering the multi-objective nature of the system it adds weight to the sub-evaluation function to build a single objective group control elevator system mathematical model and propose its constraint conditions, decision variables and objective function.Fourth, this paper recognizes the pattern of the group control elevator system by using Random Forests Algorithm. Through analysis on the data system, it gets the traffic pattern of the elevator system. Then by using the Random Forests Algorithm, it trained a group of decision trees based on the data system. Then it brings the real elevator data into the decision tree to make decision analysis and let the decision tree vote to decide the current traffic pattern.Fifth, by using the Hungarian Algorithm, this paper designs the scheduling scheme of the group control elevator system based on the traffic pattern which has been calculated out by the Random Forests Algorithm to achieve the multi-objective optimization of the group control elevator system.Sixth, through computer simulation, this paper verifies the effectiveness and superiority of the pattern recognition and scheduling scheme of the group control elevator system on the basis of the Random Forests Algorithm and Hungary Algorithm.Seventh, this paper summarizes the research progress and looks forward to further research content.