The Research Of Elevator Group Control Scheduling Policy On Robust Optimization

This thesis mainly focuses on the research of elevator group control scheduling policy on robust optimization.Robust optimization has been remarkable by many researchers as a new method for dealing with uncertain optimization problems. After having an analysis on uncertain optimization problems, the thesis introduces the research results of Ben-Tal and Bertsimas, and mainly has a detail discription on robust linear programming and robust discrete optimization. This thesis also introduces how to model the uncertain optimization problems and gets the robust optimization model, indicating the policy of prior-analysis. That is taking into the parameter uncertainy account in the optimizaiton model, making sure of constaint satisfaction on the realization of all the parameters those make values in the given uncertain set, and not ordering the parameter to obey some probability distributing. After that, the thesis mainly introduces the core problem, which is how to transform the robust optimization model into the robust counterpart. The purpose is to tranform the prior uncertain optimization problem into a tractable optimization problem, which can be solved by optimization software.With the development of the society, the requirement of a better vertical transpatation service for the people in the office building makes the elevator group control scheduling becoming an important problem. There is a change from the classical methods to artificial methods on the problem. On the basis of analyzing the methods those mentioned above, this thesis mainly discusses the difficulty of the elevator group control scheduling, which is the uncertainty of traffic flow. This thesis analyzes the elevator group control scheduling from the point of programming, incoporating with the Bertsimas research formulation on robust optimization. Traffic flow is as the uncertainty parameter. The minimizing cost of passengers' waiting time is as the objective. The robust optimization model of elevator group control scheduling is a robust 0-1 discrete optimization. At the same time, this thesis not only adopts the idea of dispatching floor and the policy of redispatching, but also thinks about the door state of the elevator during the computation of the cost of passengers' waiting time, which enhances the reasonablility of the scheduling. After the deduction on robust 0-1 discrete optimization, the robust counterpart is equal to a...