A Study On Elevator Group Control Algorithm Based On SVM

With the rapid development of modernized city construction, people's living standards are gradually improving, and also the high-rise and intelligent buildings are growing rapidly. Accordingly, as a vertical transport, the elevator is widely in used. To meet with the needs of high-rise commercial or residential buildings, the amount of elevators installed is generally increasing. In this circumstance, obviously it is really important to improve the group control of lifts, which based on the changes in buildings'traffic, as to ensure the optimum transport of buildings.However, research in the algorithm of high efficient and intelligent elevator group control system has become a hot field in building automation. Today, the scheduling level of domestic group control in elevator is not ideal yet. Actually, most advanced system of group control is made by foreign Elevator corporations. The control methods and techniques by national independent research and independent copyright are seldom put into practice. Consequently, deep study in EGCS will intensely promote the development of domestic elevator industry.The characteristics of EGCS which conclude occasionality of Elevator traffic data, randomness of call signals, and diversity of elevator group control target, pose the hard part of elevator group control. Presently, according to the modern research of EGCS, it focuses on six aspects as follow,(1)Based on the study of fuzzy logic operations, constitute fuzzy rules of the elevator group control;(2)Based on the optimum reorganization of the collection in call signal, make predictive models with indicators such as average waiting time, then make a further study in self-learning predictive control;(3) Act as an important part of the dynamic characteristics of elevator traffic, group control is put into study;(4) Introduce artificial intelligence algorithms to the elevator group control system;(5) Regard the elevator group control system an important branch of intelligent building's integrated function study;(6) Researches on elevator group control of computer network technology-based This paper starts with the key points of elevator group control, and combined with practical projects. On the basis of researches on EGCS system function, system structure, and performance indicators, it proposes the algorithm method of elevator group control which depends on support vector machines.Firstly, it described domestic and foreign situation and development trend of elevator group control algorithms. In this part, the working principle and core technology of elevator group control were studied. Secondly, analysis combined with the dynamic characteristics of elevator traffic, propose the performance requirements and performance evaluation function of EGCS. Subsequently, it adopts data of elevator traffic flow to the mathematical modeling. Using support vector machines to recognize buildings' current elevator traffic pattern, as to guide the elevator group control system adopt different models for different scheduling strategies to enhance the whole performance of system. According to the shortcomings and defect of existing algorithms, the paper combines fuzzy inference; it employed the support vector machine algorithm into group control, overcome shortcomings of fuzzy inference that lack of academic ability. By means of evaluation indicators and a variety of elevator traffic patterns, this paper optimized the dispatching strategy of EGCS. Finally, in the operating system of Windows, it used the tool of object-oriented language Visual C #, developed the simulation software of EGCS. Therefore, it achieved the dynamic visual simulation operation of optimal scheduling in elevator group. Besides, it deducted that the evaluation indicates of elevator service quality is able to verify the improvements of the algorithm in the aspects of efficiency and energy consumption.