LEARNING AGENTS IN POWER MARKETS

دانشگاه آزاد اسلامی اصفهان(خوراسگان)

منو

رکورد قبلی

رکورد بعدی

" LEARNING AGENTS IN POWER MARKETS "

Document Type	:	Latin Dissertation
Language of Document	:	English
Record Number	:	150088
Doc. No	:	ET21880
Main Entry	:	KONG-WE1 LYE
Title Proper	:	LEARNING AGENTS IN POWER MARKETS
Note	:	This document is digital این مدرک بصورت الکترونیکی می باشد
Abstract	:	In this thesis, we explore the use of software agents and learning algorithms in thesimulation of power markets. Three learning algorithms are used in the agents: geneticalgorithms, reinforcement learning, and neural networks. Experiments with learning inseller agents show that they can achieve profits close to that of a monopoly in a marketwith inelastic demand. Even in a market with elastic demand, the seller agents are able toachieve similar results. We also show that when buyer agents use learning, they are ableto prevent sellers from raising prices above competitive levels, and at the same time, ableto meet their consumption demands. The effects of the presence of a transmissionnetwork on electricity prices are also explored. Our experiments show that congestions intransmission lines can lower the market concentrations at demand buses and increase thepotential of sellers exercising market power. Finally, we show how buyer agents can useneural networks in demand schedule planning and develop a cooperation scheme forbuyer agents to help lower payments. The cooperation scheme is especially advantageousto buyer agents that are in close geographical proximity in the transmission network andexperiencing high electricity prices.Thesis Supervisor: Prof. Sarosh N. TalukdarTitle: Professor of Electrical and Computer EngineeringThesis Advisor: Prof. Bruce H. KroghTitle: Professor of Electrical and Computer Engineering,..tested for theQ1 PC1 bus cardBoth these projects mere sofixare des elopment efforts tonards contributing to dlfferentaspects of Roboucs and lZ1echatronics projects m the Controls and Roboucs Group..
Subject	:	Electericl tess
	:	برق
electronic file name	:	TL45071.pdf
Title and statement of responsibility and	:	LEARNING AGENTS IN POWER MARKETS [Thesis]