Noncooperative Game-Theoretic Study For Multi-Agent Strategies In Energy Internet

The e xisting po wer grids have e volve d around the paradigm of e nergy internet,as rene wable resource is introducing into an energy market.In partic ular,po wer system is undergoing transformation that mo ves from a vertical to flat structure and,by introducing the technology of Internet and wireless communic ation,po wer s ystem is facing an unprecedented c hallenge.The use of traditional o ptimization theory has limitations as the main decision-making process becomes fair,therefore,a no vel approac h is required to analyze the interaction between s ubjects,which is envisioned in the future energy internet.Unlike the classical optimizatio n works that often maximize or minimize a certain objective function as s ystem global value,this dissertation introduces game theory and primarily handles critical issues on data collection,trading and o ptimal management of electric energy.Also,it is expected that,this dissertation wo uld pro vide a path to e xplore the benefits and drawbacks of multi-agent strate gies via the utility functions in e very proposed mechanisms.The main contributions of this dissertation are summarized as follows:(1)To impro ve e nergy effcienecy of data transmission in po wer system,a noncooperative game was proposed for po wer allocation between multiple WSN sensor nodes.In this respect,the pro posed water-injection based algorithm is use d to maximize transmission efficiency and allocation from se nsor no des to data collection nodes(Sink no de).Based on the Hessian matrix of ne gative qualitative cross-examination it is pro ved that,the Nash e quilibrium e xists for the pro posed model and the convergence of the algorithm is analyzed.(2)For the distribution and o peration of e nergy trading,a noncooperative game with a continuo us strategy set is de velope d for competitively s ailing electricity in the consideration of interactions among source,grid and lo ad.It is pro ve d that under the condition of a no nline ar utility function,there exists a pure strategy Nas h equilibrium in the proposed mo del,in which eac h s upplier can re ach to such a Nas h equilibrium by best response.The mathematic al analysis and simulation results have been provided.(3)For the participating users i n demand response,load s hifting is proposed with a re al-time electricity pricing mechanism.The so-called Fictitious Play algorithm is use d to derive the user participation in a noncooperative game of mixed strategy Nas h e quilibrium(approximately),and gives the impro ve d algorithm to adapt to the size of the main bo dy of the virtual action and clustering analysis algorithm to impro ve the algorithm converge nce speed and scale.Simulations sho w users' participating frequency so as to predict how users shift load in practice.(4)To further explore the game-therotic application in e nergy manageme nt,this dissertation introduces dynamic e volutionary game theory for designing the e nergy requirement as ked by micro-source users.Based on the re plication dynamic equation,the users' strate gy o n distributed po wer purchase will be iterate d to response to the competition between them.To solve the problem of e nergy allocations o n both sides of supply and demand,the proposed approach impro ves user s atisfaction and increases monetary benefits in micro-source.In a summary,this dissertation studies the interactions between grid multiple age nts using noncooperative game theory.In particular,we s ho w that,such a game-theoretic method c an be used to forecast the s ystem states,as well as to pro vide an effective feedback from end nodes as an optimization designing mechanism.