2011PSCE0279 1
Abstract—Presently, Independent System Operators (ISOs) in
deregulated electricity markets in the U.S. use an auction method
that minimizes the total bid cost when determining units to be on
and their generation levels (Bid Cost Minimization or BCM). It
has recently been shown that this method of auction does not
provide minimal consumer payment costs for a given set of bids
under the Market Clearing Price (MCP)-based or congestion-
dependent Locational Marginal Price (LMP) scheme. Instead, an
alternative auction method that directly minimizes the total
consumer payment (Payment Cost Minimization or PCM)
provides the payment-minimizing auction selection. Though the
use of PCM minimizes consumer payments for a given set of bids
for a day, it has not been fully illustrated that PCM minimizes
payments over a longer time period with intelligent bidders that
can adjust to the new auction mechanism. To address this
important economic issue, a novel discrete game theoretic
method of bidder behavior is used to model the competitive
nature of generating companies in the day-ahead market.
Numerical testing results show that PCM significantly reduces
consumer payments with intelligent bidders. Simulation using a
market simulator is also presented with similar results. Finally,
insight into the potential benefits of PCM is briefly presented.
Index Terms – Payment cost minimization (PCM), bid cost
minimization (BCM), electricity auction, Nash equilibrium,
strategic behavior, multi-agent simulation.
I. INTRODUCTION
HE Independent System Operators (ISOs) that oversee
wholesale electricity markets in the U.S. use an auction
mechanism to select supply and demand bids in the day-ahead
market for energy and ancillary services. Subsequently, a
pricing algorithm is executed to determine Market Clearing
Prices (MCPs) or congestion-dependent Locational Marginal
Prices (LMPs). Presently, ISOs employ the “bid cost
minimization” (BCM) auction method that minimizes the total
bid cost given the set of bids. However, since generating
companies may not bid according to their true production
costs, bid cost minimization may not always equate to social
welfare maximization.
Manuscript received September 30, 2010. This work was supported in part
by the National Science Foundation under grants ECS-0621936 and ECCS-
1028870.
D. P. Ghosh was with the Department of Electrical and Computer
Engineering, University of Connecticut, and is now with the School of
Electrical and Computer Engineering, Cornell University, Ithaca, NY 14853
USA (Phone: 607-255-4346; e-mail: dpg65@cornell.edu).
P. B. Luh is with the Department of Electrical and Computer
Engineering, University of Connecticut, Storrs, CT 06269-2157 USA (Phone:
860-486-4821; e-mail: peter.luh@uconn.edu).
It has been shown recently that an alternative method of
auction selection, known as “payment cost minimization”
(PCM), results in a lower cost of energy for consumers than
BCM does for a given set of supply bids for a day, where
demand is assumed given for simplicity. This is because the
MCP or LMP settlement scheme is not consistent with the
BCM auction mechanism [1, 2].
However, market participants may behave differently when
the PCM auction mechanism is used. Thus, it is important to
determine if PCM can deliver lower energy prices consistently
given the intelligent and competitive nature of generating
companies. It has been shown that PCM can do this for a 24-
hour matrix-based game theoretic model with the Nash
equilibrium as the solution concept [3]. In the present study, it
will be shown that the Nash equilibria of the PCM game
consistently specify a strategy tuple that provides lower
energy costs for consumers over a 14-day period.
The effects of BCM and PCM are also studied using an
electricity market simulator, EMCAS [4]. A four-supplier,
multiple-day simulation was constructed with scaled real
loads. It was observed that PCM lowered energy costs for
consumers. Key improvements in this study over previous
ones include the use of simulation techniques such as EMCAS
and a mixed integer programming package, CPLEX, to
execute the BCM and PCM auction selections.
II. LITERATURE REVIEW
Many studies analyzing BCM exist in the literature.
Generally, the Nash equilibrium is used as a solution concept
for bid and auction simulations. Many studies focus on
continuous games as they reflect ISO markets. Further, [5, 6]
use mathematical programming methods with solutions
derived from the Karush-Kuhn-Tucker conditions. Due to the
difficulty of modeling continuous games and the fact that
important discrete unit startup and shut down decisions cannot
be considered in such formulations, later studies employ a
discrete game theoretic method, in which bid levels are
discretized at several intervals, and Nash equilibria are
searched among the strategy tuples.
Meanwhile, few studies of PCM exist. Zhao et. al.
employed a method of bid discretization and a novel algorithm
for the equilibrium search while considering discrete unit
startup and shut down decisions [7]. This method provides
limited results on the potential long-term effects of PCM
implementation. Here, a 14-day period is simulated using a
similar method of discrete bidding. Additionally, a 30-day
period is simulated using EMCAS.
T
Analysis and Simulation of Payment Cost Minimization
and Bid Cost Minimization with Strategic Bidders
Dipayan P. Ghosh and Peter B. Luh, Fellow, IEEE
978-1-61284-788-7/11/$26.00 ©2011 IEEE