Font Size: a A A

Optimal Control Of Multi-agent Systems Allowing Distributed Choices

Posted on:2012-08-10Degree:MasterType:Thesis
Country:ChinaCandidate:Z C LiuFull Text:PDF
GTID:2178330335455427Subject:Control theory and control engineering
Abstract/Summary:PDF Full Text Request
Recently, multi-agent systems and related techniques, owing to their advantages, supply an effective method for several complicated problems and thus receive extensive interest from various disciplines. Especially, for the innovativeness and widely-applied contexts, researches on cooperative control systems involving multiple agents allowing distributed choices come out and are in the ascendant. And this opens a new horizon in the cross-area of information and control theory.The present thesis proposes and solves time-energy optimal control problems for a class of linear multi-agent systems with distributed choices. This research is based on current work on multi-agent systems and starts with reviewing literatures referring to relevant investigation problems. Main research work of this thesis and major contributions are listed as below:First of all, a basic mathematical model for choice-effected multi-agent systems are constructed along with the introduction of some related concepts. In order to solve the time-energy optimal control problem for the system, applying the approach of classical calculus of variation, this thesis deduces necessary conditions of optimality subject to fixed terminal states and free terminal time which can minimize a composite cost function.To simplify the problem and facilitate solving, scalar systems are studied first. Distributed optimal controls and analytical solutions to minimum terminal time are derived for two-agent systems and multiple ones respectively according to necessary conditions of optimality. Afterward, extension is made to the optimal control of vector systems. Based on the controllability analysis of multi-agent systems, optimal control problem for two-agent vector systems that is completely controllable is solved first. For a class of partial controllable systems, the author proposes a decoupling method which can translate a two-agent system into two controllable subsystems. Therefore, optimal control solutions and optimal terminal time solutions are available for both types of systems and control results are compared. In the thesis, theoretical analysis is made for the results obtained in each section, so as to disclose the specialty of distributed choice-effected systems and demonstrate the structure of control sets, optimal terminal time and control cost. In addition, corresponding simulations show the control effects of the designed distributed strategies and validate the correctness of optimal time solutions simultaneously.Finally, the author concludes all the work completed in the present thesis and foresees the research prospect of multi-agent systems allowing distributed choices.
Keywords/Search Tags:Multi-agent System, Distributed Choices, Time-Energy Optimal, Cooperative Control
PDF Full Text Request
Related items