On Robust Optimization Model Of Uncertain Linear Systems

The problem of optimization scheduling of uncertain systems has been researched extensively due to its high academic and practical significance. On the base of deep analysis of uncertain optimization problems, the dissertation mainly researches these following contents: firstly summarizing the internationally researching situation of robust optimization, and secondly studying questions such as how to model and how to transform, aiming at uncertain linear systems, and then perfecting primal-dual interior-point method to acquire its modeling solution. Finally, the robust optimization modeling technique is applied in the problem of elevator group scheduling in order to supply a new method for the uncertain problem of traffic flows.After the deep analysis of the origin of kinds of uncertain factors, the dissertation expatiates from two different ways such as traditional uncertain optimization (for example, fuzzy optimization or stochastic optimization) and robust optimization. It also lucubrates the topic, difficulties, methods and application of robust optimization, which is separated into four different types whose basic principle are introduced individually, and forecasts the future direction audaciously. Afterwards, along the steps of the famous master of Ben-Tal and Nemirovski, according to the characteristic of uncertain linear systems, the dissertation researches the determination of ellipsoidal uncertain sets, the establishment of robust optimization model, the transformation of robust counterpart, and the discussion of computation tractability. With an eye to the importance of interior-point method in resolving optimization problem, the dissertation complements the deficiency of how to determinate original point in reference, and makes it more perfect.As a new way to resolve uncertain optimization scheduling problems, it should be the first time to introduce robust optimization into disposing elevator group scheduling problem, which is a kind of large-scale dynamical optimization problem. On the base of detailed analysis of elevator group scheduling field, three difficulties during elevator group scheduling are summarized, namely traffic flow forecast, objective function and uncertain problem of scheduling. Combined with the former research results on elevator group scheduling problem, the principle of robust...