Upload
williamb285
View
224
Download
0
Embed Size (px)
Citation preview
7/30/2019 Competitive Equilibria for Stochastic Dynamic Markets
1/90
Dynamics of Power*LCCC Lund University, Sweden
Based on tutorial & panel lectures at Energy Systems Week, Cambridge UK
Sean P. Meyn
Joint work with: In-Koo Cho, Anupama Kowli, Matias Negrete-Pincetic,Ehsan Shafieeporfaard, Uday Shanbhag, and Gui Wang
Coordinated Science Laboratoryand the Department of Electrical and Computer Engineering
University of Illinois at Urbana-Champaign, USA
Thanks to NSF, AFOSR, and DOE / TCIPG
*Dynamik i Makten (i grossistledet elmarknaderna)
May 19, 2011
http://www.newton.ac.uk/programmes/SCS/scsw07p.htmlhttp://www.newton.ac.uk/programmes/SCS/scsw07p.htmlhttp://find/http://goback/http://www.newton.ac.uk/programmes/SCS/scsw07p.html7/30/2019 Competitive Equilibria for Stochastic Dynamic Markets
2/90
Outline
1 Can You Spot the Competitive Equilibrium?
2 Competitive Equilibria in Dynamic Markets
3 Coping with Uncertainty and Constraints
4 Conclusions
5 References
2 / 4 0
http://find/7/30/2019 Competitive Equilibria for Stochastic Dynamic Markets
3/90
Can You Spot the Competitive Equilibrium?
Spot the Competitive Equilibrium
3 / 4 0
http://find/7/30/2019 Competitive Equilibria for Stochastic Dynamic Markets
4/90
Can You Spot the Competitive Equilibrium?
Competitive EquilibriumStandard economic setting
Perfect competition
Long-run setting with uncertainty:
KD(GD) = E
etWD(t) dtKS(GS) = E
etWS(t), dt
Consumers and suppliers each wish to maximize their individual welfare.Welfare functions defined with a price-process {P(t) : t 0}
4 / 4 0
C Y S h C i i E ilib i ?
http://find/http://goback/7/30/2019 Competitive Equilibria for Stochastic Dynamic Markets
5/90
Can You Spot the Competitive Equilibrium?
Competitive EquilibriumStandard economic setting
Perfect competition
Long-run setting with uncertainty:
KD(GD) = E
etWD(t) dtKS(GS) = E
etWS(t), dt
Consumers and suppliers each wish to maximize their individual welfare.Welfare functions defined with a price-process {P(t) : t 0}
Price-taking assumption
Key assumption of equilibrium theory: The price of power P(t) does notdepend on the decisions of the market agents.
4 / 4 0
http://find/7/30/2019 Competitive Equilibria for Stochastic Dynamic Markets
6/90
Can You Spot the Competitive Equilibrium?
7/30/2019 Competitive Equilibria for Stochastic Dynamic Markets
7/90
Can You Spot the Competitive Equilibrium?
Competitive EquilibriumEfficiency
Efficient Equilibrium
Social Planners Problem:
max K(G) = E etWS(t) + WD(t) dtSuppose that there exists a price process {P(t)} that forms anequlibrium: The consumers and suppliers agree, GS = GD = G.
Suppose the agreed upon decisions G solve the SPP.
Then, the market is called efficient.
5 / 4 0
Can You Spot the Competitive Equilibrium?
http://find/7/30/2019 Competitive Equilibria for Stochastic Dynamic Markets
8/90
Can You Spot the Competitive Equilibrium?
Competitive EquilibriumEfficiency
Efficient Equilibrium
Social Planners Problem:
max K(G) = E etWS(t) + WD(t) dtSuppose that there exists a price process {P(t)} that forms anequlibrium: The consumers and suppliers agree, GS = GD = G.
Suppose the agreed upon decisions G solve the SPP.
Then, the market is called efficient.
Lets look for examples of efficient equilibria!
5 / 4 0
http://find/7/30/2019 Competitive Equilibria for Stochastic Dynamic Markets
9/90
http://find/7/30/2019 Competitive Equilibria for Stochastic Dynamic Markets
10/90
http://find/http://goback/7/30/2019 Competitive Equilibria for Stochastic Dynamic Markets
11/90
Can You Spot the Competitive Equilibrium?
http://find/7/30/2019 Competitive Equilibria for Stochastic Dynamic Markets
12/90
p p q
Efficient Equilibrium?Australia, January 16, 2007
Price(Aus$/M
Wh)
Volume(MW)
Demand
Demand
Prices
Prices
24:00
23:00
22:00
21:00
20:00
19:00
18:00
17:00
16:00
15:00
14:00
13:00
12:00
11:00
10:00
09:00
08:00
07:00
06:00
05:00
04:00
03:00
02:00
01:00
00:00
24:00
23:00
22:00
21:00
20:00
19:00
18:00
17:00
16:00
15:00
14:00
13:00
12:00
11:00
10:00
09:00
08:00
07:00
06:00
05:00
04:00
03:00
02:00
01:00
00:00
19,000
10,000
1,400
1,200
1,000
800
600
400
200
0
1,000
0
- 1,000
-1,500
1,000
-1,000
9,000
8,000
10,000
7,000
6,000
5,000
4,000
3,000
2,000
1,000
0
Tasmania
Victoria
7 / 4 0
Can You Spot the Competitive Equilibrium?
http://find/http://goback/7/30/2019 Competitive Equilibria for Stochastic Dynamic Markets
13/90
Efficient Equilibrium?Australia, January 16, 2007
Price(Aus$/M
Wh)
Volume(MW)
Demand
Demand
Prices
Prices
24:00
23:00
22:00
21:00
20:00
19:00
18:00
17:00
16:00
15:00
14:00
13:00
12:00
11:00
10:00
09:00
08:00
07:00
06:00
05:00
04:00
03:00
02:00
01:00
00:00
24:00
23:00
22:00
21:00
20:00
19:00
18:00
17:00
16:00
15:00
14:00
13:00
12:00
11:00
10:00
09:00
08:00
07:00
06:00
05:00
04:00
03:00
02:00
01:00
00:00
19,000
10,000
1,400
1,200
1,000
800
600
400
200
0
1,000
0
- 1,000
-1,500
1,000
-1,000
9,000
8,000
10,000
7,000
6,000
5,000
4,000
3,000
2,000
1,000
0
Tasmania
Victoria
Fires cause chaos Is this efficient?
7 / 4 0
Can You Spot the Competitive Equilibrium?
http://find/7/30/2019 Competitive Equilibria for Stochastic Dynamic Markets
14/90
Efficient Equilibrium?Midwest ISO today: Friday afternoon, March 4, 2011 3:30 p.m.
-2000.00
8 / 4 0
Can You Spot the Competitive Equilibrium?
http://find/7/30/2019 Competitive Equilibria for Stochastic Dynamic Markets
15/90
Efficient Equilibrium?Midwest ISO today: Friday afternoon, March 4, 2011 3:50 p.m.
-762.55
9 / 4 0
Can You Spot the Competitive Equilibrium?
http://find/7/30/2019 Competitive Equilibria for Stochastic Dynamic Markets
16/90
Efficient Equilibrium?Midwest ISO today: Friday afternoon, March 4, 2011 4:15 p.m.
-1881.07
10/40
Can You Spot the Competitive Equilibrium?
http://find/7/30/2019 Competitive Equilibria for Stochastic Dynamic Markets
17/90
Efficient Equilibrium?Midwest ISO today: Friday afternoon, March 4, 2011 4:30 p.m.
-115.68
11/40
Can You Spot the Competitive Equilibrium?
http://find/7/30/2019 Competitive Equilibria for Stochastic Dynamic Markets
18/90
Efficient Equilibrium?Texas today: Winter of 2011
5am 10am 3pm 8pm500
0
1000
2000
$/MWh
February 2, 2011$/MWh
10
0
10
20
40
60
80
5am 10am 3pm 8pm
Power Prices in TexasJanuary 31, 2011
12/40
Can You Spot the Competitive Equilibrium?
http://find/7/30/2019 Competitive Equilibria for Stochastic Dynamic Markets
19/90
Efficient Equilibrium?Texas today: Winter of 2011
5am 10am 3pm 8pm500
0
1000
2000
$/MWh
February 2, 2011$/MWh
10
0
10
20
40
60
80
5am 10am 3pm 8pm
Power Prices in TexasJanuary 31, 2011
There will be multiple autopsies of the causes for the latest power breakdowns ... Who profited
off this near-meltdown and what can be done to incentivize power producers to maintainadequate reserve capacity for emergencies rather than waiting for emergency windfalls? HOUSTON CHRONICLE, Feb 12, 2011
New report hits ERCOT, electricity deregulation: A report released Monday concludes thatelectric deregulation has cost Texas residential consumers more than $11 billion in higherrates... Dallas Morning News, Feb 14, 2011
12/40
Can You Spot the Competitive Equilibrium?
http://find/7/30/2019 Competitive Equilibria for Stochastic Dynamic Markets
20/90
Efficient Equilibrium?Texas today: Winter of 2011
What do theexperts say?
5am 10am 3pm 8pm500
0
1000
2000
000$/MWh
February 2, 2011$/MWh
10
0
10
20
40
60
80
5am 10am 3pm 8pm
Power Prices in TexasJanuary 31, 2011
13/40
Can You Spot the Competitive Equilibrium?
http://find/7/30/2019 Competitive Equilibria for Stochastic Dynamic Markets
21/90
Efficient Equilibrium?Texas today: Winter of 2011
What do theexperts say?
5am 10am 3pm 8pm500
0
1000
2000
000$/MWh
February 2, 2011$/MWh
10
0
10
20
40
60
80
5am 10am 3pm 8pm
Power Prices in TexasJanuary 31, 2011
...long-term contracts could reduce market power... ...thelarger the proportion of total demand auctioned in advance, the lower areboth the contract and the average spot price of energy...
M. Soledad Arellano and Pablo Serra, 2010
13/40
Can You Spot the Competitive Equilibrium?
http://find/7/30/2019 Competitive Equilibria for Stochastic Dynamic Markets
22/90
Efficient Equilibrium?Texas today: Winter of 2011
What do theexperts say?
5am 10am 3pm 8pm500
0
1000
2000
000$/MWh
February 2, 2011$/MWh
10
0
10
20
40
60
80
5am 10am 3pm 8pm
Power Prices in TexasJanuary 31, 2011
...long-term contracts could reduce market power... ...thelarger the proportion of total demand auctioned in advance, the lower areboth the contract and the average spot price of energy...
M. Soledad Arellano and Pablo Serra, 2010
...forward markets do not mitigate market power... ...forwardmarkets systematically enhance market power in some symmetriccapacity-constrained markets...
Frederic Murphy and Yves Smeers, 2010
13/40
Can You Spot the Competitive Equilibrium?
http://find/7/30/2019 Competitive Equilibria for Stochastic Dynamic Markets
23/90
Efficient Equilibrium?New Zealand today: March 25, 2011
A typical day in the New Zealand power market on the N. Island
Stratford
Otahuhu
http://www.electricityinfo.co.nz/
0
50
100
Nodal Power Prices in NZ: $/MWh
4am 9am 2pm 7pm
14/40
Can You Spot the Competitive Equilibrium?
http://find/http://goback/7/30/2019 Competitive Equilibria for Stochastic Dynamic Markets
24/90
Efficient Equilibrium?New Zealand today: March 26, 2011
$25 million dollars extracted by the generators in just six hours
Stratford
Otahuhu
http://www.electricityinfo.co.nz/4am 9am 2pm 7pm0
10,000
20,000
Nodal Power Prices in NZ: $/MWh
15/40
Can You Spot the Competitive Equilibrium?
http://find/7/30/2019 Competitive Equilibria for Stochastic Dynamic Markets
25/90
Efficient Equilibrium?New Zealand today: March 26, 2011
$25 million dollars extracted by the generators in just six hours
Stratford
Otahuhu
http://www.electricityinfo.co.nz/4am 9am 2pm 7pm0
10,000
20,000
Nodal Power Prices in NZ: $/MWh
Efficient? Energy consultant Bryan Leyland said the high wholesale prices showed howdysfunctional the electricity market is.
Jacking the prices up sends no worthwhile signal to anyone it is nothing to do with ashortage of generating capacity, he said. It just exposes the nonsense of it all.
15/40
Can You Spot the Competitive Equilibrium?
http://find/7/30/2019 Competitive Equilibria for Stochastic Dynamic Markets
26/90
Efficient Equilibrium?New Zealand today: March 26, 2011
$25 million dollars extracted by the generators in just six hours
Stratford
Otahuhu
http://www.electricityinfo.co.nz/4am 9am 2pm 7pm
0
10,000
20,000
Nodal Power Prices in NZ: $/MWh
Efficient! Preliminary view of NZ Electrical Authority: Genesis was not guiltyof manipulative, ... or deceptive conduct.
16/40
Can You Spot the Competitive Equilibrium?
http://find/7/30/2019 Competitive Equilibria for Stochastic Dynamic Markets
27/90
Efficient Equilibrium?New Zealand today: March 26, 2011
$25 million dollars extracted by the generators in just six hours
Stratford
Otahuhu
http://www.electricityinfo.co.nz/4am 9am 2pm 7pm
0
10,000
20,000
Nodal Power Prices in NZ: $/MWh
Efficient! Preliminary view of NZ Electrical Authority: Genesis was not guiltyof manipulative, ... or deceptive conduct. However, high prices threatened to
undermine confidence in, and ... damage the integrity and reputation of the
wholesale electricity market 3:59 PM Friday May 6, 2011 www.nzherald.co.nz
16/40
Can You Spot the Competitive Equilibrium?
http://goforward/http://find/http://goback/7/30/2019 Competitive Equilibria for Stochastic Dynamic Markets
28/90
Efficient Equilibrium?Its not just about power!
Retail prices o onions in IndiaIndian Rupee per kg
Source: Ministry o Consumer Afairs
Jan Mar May July Sep Nov Jan Mar May July Sep Nov Jan
40
50
30
10
-10
20
0
Chennai
Delhi
Mumbai
Patna
2009 2010 2011
Delhi
Mumbai
Rs. 60
Rs. 55
Rs. 50Rs. 44
Rs. 37Rs. 43
Rs. 25
Rs. 13
Rs. 14
Rs. 9
Rs. 15.5
Rs. 29
Dec 20 Dec 27 Jan 10 Jan 24 Feb 7 Feb 11
Sharp Rise & Sharp Fall
Indias ban on onion exports to Nepal ... caused
wholesale price to jump 50% in a week. NepaliEconomy.com, Jan 3, 2011
Soaring onion prices pushed food inflation
again to a double-digit mark... Indian Express, Feb 12, 2011
In India, onion prices are as politically sensitive as mortgage rates are inAustralia. ... rising cost of the staple has helped change governments inIndia. Fidelity International, Jan 2011
17/40
Competitive Equilibria in Dynamic Markets
http://goforward/http://find/http://goback/7/30/2019 Competitive Equilibria for Stochastic Dynamic Markets
29/90
Competitive Equilibria in Dynamic Markets
18/40
Competitive Equilibria in Dynamic Markets
El i i M k T d
http://find/7/30/2019 Competitive Equilibria for Stochastic Dynamic Markets
30/90
Electricity Markets TodayTwo coupled markets
Day-ahead market (DAM):
Cleared one day prior to the production and delivery of energy: The ISOgenerates a schedule of generators to supply specific levels of power foreach hour over the next 24 hour period.
Facilitates the scheduling of generating units, and allows for hedging against
uncertainty
19/40
Competitive Equilibria in Dynamic Markets
El i i M k T d
http://find/7/30/2019 Competitive Equilibria for Stochastic Dynamic Markets
31/90
Electricity Markets TodayTwo coupled markets
Day-ahead market (DAM):
Cleared one day prior to the production and delivery of energy: The ISOgenerates a schedule of generators to supply specific levels of power foreach hour over the next 24 hour period.
Facilitates the scheduling of generating units, and allows for hedging against
uncertainty
Real-time market (RTM):
As supply and demand are not perfectly predictable, the RTM plays therole of fine-tuning this resource allocation process
RTM is the focus here
19/40
Competitive Equilibria in Dynamic Markets
RTM M d l
http://find/7/30/2019 Competitive Equilibria for Stochastic Dynamic Markets
32/90
RTM ModelDynamic model for reserves
Cho & Meyn model: Math model explains volatile prices in power markets. SIAM News, Robinson. 2005.
R(t) = Available power Demand = G(t)D(t)
D(t) = Actual demand ForecastFor computation: Deviation in demand D is modeled as Brownian motion
G(t): Deviation in on-line capacity from day-ahead market
20/40
Competitive Equilibria in Dynamic Markets
RTM M d l
http://find/7/30/2019 Competitive Equilibria for Stochastic Dynamic Markets
33/90
RTM ModelDynamic model for reserves
Cho & Meyn model: Math model explains volatile prices in power markets. SIAM News, Robinson. 2005.
R(t) = Available power Demand = G(t)D(t)
D(t) = Actual demand ForecastFor computation: Deviation in demand D is modeled as Brownian motion
G(t): Deviation in on-line capacity from day-ahead market
Economic Friction
Generation cannot increase instantaneously:
For all t 0 and t > t, G(t
)G(t)t t
20/40
Competitive Equilibria in Dynamic Markets
RTM M d l
http://find/7/30/2019 Competitive Equilibria for Stochastic Dynamic Markets
34/90
RTM ModelDynamic model for reserves
Cho & Meyn model: Math model explains volatile prices in power markets. SIAM News, Robinson. 2005.
R(t) = Available power Demand = G(t)D(t)
D(t) = Actual demand ForecastFor computation: Deviation in demand D is modeled as Brownian motion
G(t): Deviation in on-line capacity from day-ahead market
Economic Friction
Generation cannot increase instantaneously:
For all t 0 and t > t, G(t
)G(t)t t
In recent work we also impose lower bounds on generation, as well asnetwork constraints.
20/40
Competitive Equilibria in Dynamic Markets
Market Analysis: A beautiful world
http://find/http://goback/7/30/2019 Competitive Equilibria for Stochastic Dynamic Markets
35/90
Market Analysis: A beautiful world...Perfect competition
Dynamic market equilibria under the most ideal circumstances:
21/40
Competitive Equilibria in Dynamic Markets
Market Analysis: A beautiful world
http://find/7/30/2019 Competitive Equilibria for Stochastic Dynamic Markets
36/90
Market Analysis: A beautiful world...Perfect competition
Dynamic market equilibria under the most ideal circumstances:
Price manipulation is excluded
21/40
Competitive Equilibria in Dynamic Markets
Market Analysis: A beautiful world
http://find/7/30/2019 Competitive Equilibria for Stochastic Dynamic Markets
37/90
Market Analysis: A beautiful world...Perfect competition
Dynamic market equilibria under the most ideal circumstances:
Price manipulation is excludedNo ENRON games no market power
21/40
Competitive Equilibria in Dynamic Markets
Market Analysis: A beautiful world
http://find/7/30/2019 Competitive Equilibria for Stochastic Dynamic Markets
38/90
Market Analysis: A beautiful world...Perfect competition
Dynamic market equilibria under the most ideal circumstances:
Price manipulation is excludedNo ENRON games no market power
Externalities are disregardedNo government mandates
21/40
Competitive Equilibria in Dynamic Markets
Market Analysis: A beautiful world
http://find/7/30/2019 Competitive Equilibria for Stochastic Dynamic Markets
39/90
Market Analysis: A beautiful world...Perfect competition
Dynamic market equilibria under the most ideal circumstances:
Price manipulation is excludedNo ENRON games no market power
Externalities are disregardedNo government mandates
When wind generation is included, the consumers own and operatethese resources
21/40
Competitive Equilibria in Dynamic Markets
Market Analysis: A beautiful world
http://find/7/30/2019 Competitive Equilibria for Stochastic Dynamic Markets
40/90
Market Analysis: A beautiful world...Perfect competition
Dynamic market equilibria under the most ideal circumstances:
Price manipulation is excludedNo ENRON games no market power
Externalities are disregardedNo government mandates
When wind generation is included, the consumers own and operatethese resources
What does an efficient equilibrium look like?
21/40
Competitive Equilibria in Dynamic Markets
Market Analysis
http://find/7/30/2019 Competitive Equilibria for Stochastic Dynamic Markets
41/90
Market AnalysisSecond Welfare Theorem
Efficient Equilibrium
max K(G) = E
etWS(t) + WD(t)
dt
.
s.t. GS(t) = GD(t) for all t
22/40
Competitive Equilibria in Dynamic Markets
Market Analysis
http://find/7/30/2019 Competitive Equilibria for Stochastic Dynamic Markets
42/90
Market AnalysisSecond Welfare Theorem
Efficient Equilibrium
max K(G) = E
etWS(t) + WD(t)
dt
.
s.t. GS(t) = GD(t) for all t
Special case: Welfare functions are piecewise linear,
WS(t) := P(t)GS(t) cGS(t)
WD(t) := v min(D(t), GD(t))
cbo max(0,RD(t)) P(t)GD(t)
22/40
Competitive Equilibria in Dynamic Markets
Market Analysis
http://find/7/30/2019 Competitive Equilibria for Stochastic Dynamic Markets
43/90
Market AnalysisSecond Welfare Theorem
Efficient Equilibrium
max K(G) = E
etWS(t) + WD(t)
dt
s.t. GS(t) = GD(t) for all t
Key component of equilibrium theory: Perfect competition
Price-taking assumption:The price of power P(t) in the RTM is assumed to be exogenous (it doesnot depend on the decisions of the market agents).
23/40
Competitive Equilibria in Dynamic Markets
Market Analysis
http://find/7/30/2019 Competitive Equilibria for Stochastic Dynamic Markets
44/90
Market AnalysisSecond Welfare Theorem
Second Welfare Theorem Lagrangian Decomposition
max K(G) = maxGS
E
etWS(t) + (t)GS(t)
dt
+ maxGD E
etWD(t) (t)GD(t)
dt
24/40
Competitive Equilibria in Dynamic Markets
Market Analysis
http://find/7/30/2019 Competitive Equilibria for Stochastic Dynamic Markets
45/90
ySecond Welfare Theorem
Second Welfare Theorem Lagrangian Decomposition
max K(G) = maxGS
E
etWS(t) + (t)GS(t)
dt
+ maxGD E
etWD(t) (t)GD(t)
dt
Assume: Social planners problem has a solution,and there is no duality gap.
24/40
Competitive Equilibria in Dynamic Markets
Market Analysis
http://find/7/30/2019 Competitive Equilibria for Stochastic Dynamic Markets
46/90
ySecond Welfare Theorem
Second Welfare Theorem Lagrangian Decomposition
max K(G) = maxGS
E
etWS(t) + (t)GS(t)
dt
+ maxGD E
etWD(t) (t)GD(t)
dt
Assume: Social planners problem has a solution,and there is no duality gap.
Then P(t) = P(t) + (t) provides an efficient equilibrium.
24/40
Competitive Equilibria in Dynamic Markets
Market Analysis
http://find/7/30/2019 Competitive Equilibria for Stochastic Dynamic Markets
47/90
ySecond Welfare Theorem
Second Welfare Theorem Lagrangian Decomposition
max K(G) = maxGS
E
etWS(t) + (t)GS(t)
dt
+ maxGD E
etWD(t) (t)GD(t)
dt
Assume: Social planners problem has a solution,and there is no duality gap.
Then P(t) = P(t) + (t) provides an efficient equilibrium.
What does an efficient equilibrium look like?
24/40
Competitive Equilibria in Dynamic Markets
What does an efficient equilibrium look like?
http://find/7/30/2019 Competitive Equilibria for Stochastic Dynamic Markets
48/90
qAnswer: Marginal value
Equilibrium price
The equilibrium price process is a function of equilibrium reserves:
P(t) = p(Re(t))
25/40
Competitive Equilibria in Dynamic Markets
What does an efficient equilibrium look like?
http://find/7/30/2019 Competitive Equilibria for Stochastic Dynamic Markets
49/90
Answer: Marginal value
Equilibrium price
The equilibrium price process is a function of equilibrium reserves:
P(t) = p(Re(t))
P(t) = p(Re(t)) is the marginal value of power to the consumer
25/40
Competitive Equilibria in Dynamic Markets
What does an efficient equilibrium look like?
http://find/7/30/2019 Competitive Equilibria for Stochastic Dynamic Markets
50/90
Answer: Marginal value
Equilibrium price
The equilibrium price process is a function of equilibrium reserves:
P(t) = p(Re(t))
P(t) = p(Re(t)) is the marginal value of power to the consumer
For linear cost/utility, marginal value is piecewise constant,
p(re) = (v + cbo)I{re < 0}
25/40
Competitive Equilibria in Dynamic Markets
Market Equilibrium
http://find/7/30/2019 Competitive Equilibria for Stochastic Dynamic Markets
51/90
Price dynamics
v + cbo
0
r
Prices
Reserves
Normalized demand
P(t) = p(Re(t)): The marginal value of power to the consumer
26/40
Competitive Equilibria in Dynamic Markets
Market Equilibrium
http://find/7/30/2019 Competitive Equilibria for Stochastic Dynamic Markets
52/90
Price dynamics
v + cbo
0
r
Prices
Reserves
Normalized demand
P(t) = p(Re(t)): The marginal value of power to the consumer
Smoother prices obtained when cost/utility are strictly convex
26/40
Competitive Equilibria in Dynamic Markets
Familiar, right?
http://find/7/30/2019 Competitive Equilibria for Stochastic Dynamic Markets
53/90
Real-world price dynamics
27/40
Competitive Equilibria in Dynamic Markets
Sustainable business?
http://find/7/30/2019 Competitive Equilibria for Stochastic Dynamic Markets
54/90
Marginal value of electricitymay be
v + cbo
= $100,000/MWh!
28/40
Competitive Equilibria in Dynamic Markets
Sustainable business?
http://find/7/30/2019 Competitive Equilibria for Stochastic Dynamic Markets
55/90
Marginal value of electricitymay be
v + cbo
= $100,000/MWh!
However,
28/40
Competitive Equilibria in Dynamic Markets
Sustainable business?
http://find/7/30/2019 Competitive Equilibria for Stochastic Dynamic Markets
56/90
Marginal value of electricitymay be
v + cbo
= $100,000/MWh!
However,
Theorem 2: In this equilibrium, the average price isprecisely the marginal cost c.Proof: Lagrangian relaxation of initial condition.
28/40
Competitive Equilibria in Dynamic Markets
Sustainable business?
http://find/7/30/2019 Competitive Equilibria for Stochastic Dynamic Markets
57/90
Marginal value of electricitymay be
v + cbo
= $100,000/MWh!
However,
Theorem 2: In this equilibrium, the average price isprecisely the marginal cost c.Proof: Lagrangian relaxation of initial condition.
Is this a sustainable business?
28/40
Coping with Uncertainty and Constraints
http://find/7/30/2019 Competitive Equilibria for Stochastic Dynamic Markets
58/90
Coping with Uncertainty and Constraints
29/40
Coping with Uncertainty and Constraints
Network ConstraintsEntropic prices
http://find/http://goback/7/30/2019 Competitive Equilibria for Stochastic Dynamic Markets
59/90
Entropic prices
What is marginal value?
It is not always obvious. With the introduction of network constraints,
30/40
Coping with Uncertainty and Constraints
Network ConstraintsEntropic prices
https://netfiles.uiuc.edu/meyn/www/spm_files/MarketsIFAC2011/MarketsIFAC2011.htmlhttp://find/7/30/2019 Competitive Equilibria for Stochastic Dynamic Markets
60/90
Entropic prices
What is marginal value?
It is not always obvious. With the introduction of network constraints,
Prices can go well beyond marginal value
Prices can go well below zero See [Wang et. al., 2011]
30/40
Coping with Uncertainty and Constraints
Network ConstraintsEntropic prices
https://netfiles.uiuc.edu/meyn/www/spm_files/MarketsIFAC2011/MarketsIFAC2011.htmlhttps://netfiles.uiuc.edu/meyn/www/spm_files/MarketsIFAC2011/MarketsIFAC2011.htmlhttp://find/7/30/2019 Competitive Equilibria for Stochastic Dynamic Markets
61/90
Entropic prices
What is marginal value?
It is not always obvious. With the introduction of network constraints,
Prices can go well beyond marginal value
Prices can go well below zero See [Wang et. al., 2011]
Without price-caps, Australia might look like an efficient equilibrium:
Price(Aus$/
MWh)
Price(Aus$/M
Wh)
Volume(M
W)
Volume(MW)
Demand
Demand
Prices
Prices
24:00
23:00
22:00
21:00
20:00
19:00
18:00
17:00
16:00
15:00
14:00
13:00
12:00
11:00
10:00
09:00
08:00
07:00
06:00
05:00
04:00
03:00
02:00
01:00
00:00
24:00
23:00
22:00
21:00
20:00
19:00
18:00
17:00
16:00
15:00
14:00
13:00
12:00
11:00
10:00
09:00
08:00
07:00
06:00
05:00
04:00
03:00
02:00
01:00
00:00
19,000
10,000
1,400
1,200
1,000
800
600
400
200
0
1,000
0
- 1,000
-1,500
1,000
-1,000
9,000
8,000
10,000
7,000
6,000
5,000
4,000
3,000
2,000
1,000
0
TasmaniaVictoria
30/40
Coping with Uncertainty and Constraints
The Value of Volatile ResourcesWhat is the impact of wind solar or tidal generation?
https://netfiles.uiuc.edu/meyn/www/spm_files/MarketsIFAC2011/MarketsIFAC2011.htmlhttps://netfiles.uiuc.edu/meyn/www/spm_files/MarketsIFAC2011/MarketsIFAC2011.htmlhttp://find/7/30/2019 Competitive Equilibria for Stochastic Dynamic Markets
62/90
What is the impact of wind, solar, or tidal generation?
Requires consideration of coupled RTM and DAM. In summary:
31/40
Coping with Uncertainty and Constraints
The Value of Volatile ResourcesWhat is the impact of wind solar or tidal generation?
https://netfiles.uiuc.edu/meyn/www/spm_files/MarketsIFAC2011/MarketsIFAC2011.htmlhttp://find/7/30/2019 Competitive Equilibria for Stochastic Dynamic Markets
63/90
What is the impact of wind, solar, or tidal generation?
Requires consideration of coupled RTM and DAM. In summary:
cv0 0.1 0.2 0.3 0.4
k = 0
k = 5
k = 10
k = 15
k = 20Optimal Reserve Level
0
20
40
60
80
100
120
140x 10
3
Optimal reserves rise:
Optimal reserves increase with in-creasing penetration k, or coefficientof variation cv.
31/40
Coping with Uncertainty and Constraints
The Value of Volatile ResourcesWhat is the impact of wind, solar, or tidal generation?
https://netfiles.uiuc.edu/meyn/www/spm_files/MarketsIFAC2011/MarketsIFAC2011.htmlhttps://netfiles.uiuc.edu/meyn/www/spm_files/MarketsIFAC2011/MarketsIFAC2011.htmlhttp://find/7/30/2019 Competitive Equilibria for Stochastic Dynamic Markets
64/90
What is the impact of wind, solar, or tidal generation?
Requires consideration of coupled RTM and DAM. In summary:
cv0 0.1 0.2 0.3 0.4
k = 0
k = 5
k = 10
k = 15
k = 20Optimal Reserve Level
0
20
40
60
80
100
120
140x 10
3
Optimal reserves rise:
Optimal reserves increase with in-creasing penetration k, or coefficientof variation cv.
Social welfare falls with wind.
[Value of Volatile Resources, CDC, 2010]
31/40
Coping with Uncertainty and Constraints
The Value of Volatile ResourcesDistribution of welfare (consumers command wind resources)
https://netfiles.uiuc.edu/meyn/www/spm_files/MarketsIFAC2011/MarketsIFAC2011.htmlhttps://netfiles.uiuc.edu/meyn/www/spm_files/MarketsIFAC2011/MarketsIFAC2011.htmlhttps://netfiles.uiuc.edu/meyn/www/spm_files/MarketsIFAC2011/MarketsIFAC2011.htmlhttp://find/7/30/2019 Competitive Equilibria for Stochastic Dynamic Markets
65/90
( )
With increased volatility: Consumer welfare falls, supplier welfare rises:
0 0.1 0.2 0.3 0.40 0.1 0.2 0.3 0.4
x 106
Supplier Welfare
k = 0
k = 5
k = 10
k = 15
k = 20
x 106
Consumer Welfare
10
0
5
15
k = 0
k = 5
k = 10
k = 15k = 20
3
4
5
cv
Penetration k, coefficient of variation cv .
32/40
Conclusions
http://find/7/30/2019 Competitive Equilibria for Stochastic Dynamic Markets
66/90
Conclusions
33/40
Conclusions
Economics and KontrollDistributed control and prices
http://find/7/30/2019 Competitive Equilibria for Stochastic Dynamic Markets
67/90
The current RTM paradigm is doomed to fail:
34/40
Conclusions
Economics and KontrollDistributed control and prices
http://find/7/30/2019 Competitive Equilibria for Stochastic Dynamic Markets
68/90
The current RTM paradigm is doomed to fail:
In the competitive equilibrium nirvana:Volatile prices combined with low average-prices will halt investment.
34/40
Conclusions
Economics and KontrollDistributed control and prices
http://find/7/30/2019 Competitive Equilibria for Stochastic Dynamic Markets
69/90
The current RTM paradigm is doomed to fail:
In the competitive equilibrium nirvana:Volatile prices combined with low average-prices will halt investment.
In the real world?Strategic behavior can lead to a new crisis each year!
34/40
Conclusions
Economics and KontrollDistributed control and prices
http://find/7/30/2019 Competitive Equilibria for Stochastic Dynamic Markets
70/90
The current RTM paradigm is doomed to fail:
In the competitive equilibrium nirvana:Volatile prices combined with low average-prices will halt investment.
In the real world?Strategic behavior can lead to a new crisis each year!
However,
34/40
Conclusions
Economics and KontrollDistributed control and prices
http://find/7/30/2019 Competitive Equilibria for Stochastic Dynamic Markets
71/90
The current RTM paradigm is doomed to fail:
In the competitive equilibrium nirvana:Volatile prices combined with low average-prices will halt investment.
In the real world?Strategic behavior can lead to a new crisis each year!
However,
No one of us has all the answers!
34/40
Conclusions
Economics and KontrollDistributed control and prices
http://find/7/30/2019 Competitive Equilibria for Stochastic Dynamic Markets
72/90
The current RTM paradigm is doomed to fail:
In the competitive equilibrium nirvana:Volatile prices combined with low average-prices will halt investment.
In the real world?Strategic behavior can lead to a new crisis each year!
However,
No one of us has all the answers!
Collaboration is required Between researchers in economics, systemssciences, statistical sciences, and power engineers.
How could two smart people come to such different conclusions?I had to get to the bottom of this. MacKay 2009
34/40
Conclusions
Economics and KontrollDistributed control and prices
http://find/7/30/2019 Competitive Equilibria for Stochastic Dynamic Markets
73/90
The current RTM paradigm is doomed to fail:
In the competitive equilibrium nirvana:Volatile prices combined with low average-prices will halt investment.
In the real world?Strategic behavior can lead to a new crisis each year!
However,
No one of us has all the answers!
Collaboration is required Between researchers in economics, systemssciences, statistical sciences, and power engineers.
How could two smart people come to such different conclusions?I had to get to the bottom of this. MacKay 2009
A gentle debate. Please, no shouting!
34/40
Conclusions
Economics and KontrollEconomics for an Entropic Grid
http://find/7/30/2019 Competitive Equilibria for Stochastic Dynamic Markets
74/90
How to create policies to protect participants on both sides of the market,
while creating incentives for R&D on renewable energy?
Our community must consider long-term planning and policy, along with
traditional systems operations
35/40
Conclusions
Economics and KontrollEconomics for an Entropic Grid
http://find/7/30/2019 Competitive Equilibria for Stochastic Dynamic Markets
75/90
How to create policies to protect participants on both sides of the market,
while creating incentives for R&D on renewable energy?
Our community must consider long-term planning and policy, along with
traditional systems operations
Planning and Policy, includes Markets & Competition
35/40
Conclusions
Economics and KontrollEconomics for an Entropic Grid
http://find/7/30/2019 Competitive Equilibria for Stochastic Dynamic Markets
76/90
How to create policies to protect participants on both sides of the market,
while creating incentives for R&D on renewable energy?
Our community must consider long-term planning and policy, along with
traditional systems operations
Planning and Policy, includes Markets & Competition
Evolution?
35/40
Conclusions
Economics and KontrollEconomics for an Entropic Grid
http://find/7/30/2019 Competitive Equilibria for Stochastic Dynamic Markets
77/90
How to create policies to protect participants on both sides of the market,
while creating incentives for R&D on renewable energy?
Our community must consider long-term planning and policy, along with
traditional systems operations
Planning and Policy, includes Markets & Competition
Evolution? Too slow!
35/40
Conclusions
Economics and KontrollEconomics for an Entropic Grid
http://find/7/30/2019 Competitive Equilibria for Stochastic Dynamic Markets
78/90
How to create policies to protect participants on both sides of the market,
while creating incentives for R&D on renewable energy?
Our community must consider long-term planning and policy, along with
traditional systems operations
Planning and Policy, includes Markets & Competition
Evolution? Too slow!
What we need is Intelligent Design
35/40
Conclusions
Economics and KontrollEconomics for an Entropic Grid
http://find/7/30/2019 Competitive Equilibria for Stochastic Dynamic Markets
79/90
Real-world issues: Incorporate dynamics and uncertainty in a strategic
setting
36/40
Conclusions
Economics and KontrollEconomics for an Entropic Grid
http://find/7/30/2019 Competitive Equilibria for Stochastic Dynamic Markets
80/90
Real-world issues: Incorporate dynamics and uncertainty in a strategic
setting
Beyond efficiency: The economic notion of efficiency disregards issuessuch as reliability, emissions, public acceptance, or public safety.
36/40
Conclusions
Economics and KontrollEconomics for an Entropic Grid
http://find/7/30/2019 Competitive Equilibria for Stochastic Dynamic Markets
81/90
Real-world issues: Incorporate dynamics and uncertainty in a strategic
setting
Beyond efficiency: The economic notion of efficiency disregards issuessuch as reliability, emissions, public acceptance, or public safety.
Long-term incentives: How to create policies to protect participants on
both sides of the market, while creating incentives for R&D on renewableenergy?
36/40
Conclusions
Economics and KontrollEconomics for an Entropic Grid
http://find/7/30/2019 Competitive Equilibria for Stochastic Dynamic Markets
82/90
Real-world issues: Incorporate dynamics and uncertainty in a strategic
setting
Beyond efficiency: The economic notion of efficiency disregards issuessuch as reliability, emissions, public acceptance, or public safety.
Long-term incentives: How to create policies to protect participants on
both sides of the market, while creating incentives for R&D on renewableenergy?
Approach: Follow the example of highway engineering
Analysis of strategic behavior will be possible,
if agent behavior is suitably constrained
36/40 Conclusions
Economics and KontrollEconomics for an Entropic Grid
http://find/7/30/2019 Competitive Equilibria for Stochastic Dynamic Markets
83/90
Real-world issues: Incorporate dynamics and uncertainty in a strategic
setting
Beyond efficiency: The economic notion of efficiency disregards issuessuch as reliability, emissions, public acceptance, or public safety.
Long-term incentives: How to create policies to protect participants on
both sides of the market, while creating incentives for R&D on renewableenergy?
Approach: Follow the example of highway engineering
Analysis of strategic behavior will be possible,
if agent behavior is suitably constrained
The economic security of the region is at stake:We need well-designed lanes and speed limits in the energy highway!
36/40 Conclusions
Concluding Remarks
http://find/http://goback/7/30/2019 Competitive Equilibria for Stochastic Dynamic Markets
84/90
An Entropic Grid may emerge as a result of many of the proposed
Smart Grid initiativesDecision & control, simulation & learning provide tools that will helpto design an intelligent grid
37/40 Conclusions
Concluding Remarks
http://find/7/30/2019 Competitive Equilibria for Stochastic Dynamic Markets
85/90
An Entropic Grid may emerge as a result of many of the proposed
Smart Grid initiativesDecision & control, simulation & learning provide tools that will helpto design an intelligent grid
We must move beyond the traditional static competitive equilibrium
analysis
We must move beyond the traditional myopic goal of economicefficiency, taking into account externalities such as sustainabilityand reliability
37/40 Conclusions
Concluding Remarks
http://find/7/30/2019 Competitive Equilibria for Stochastic Dynamic Markets
86/90
An Entropic Grid may emerge as a result of many of the proposed
Smart Grid initiativesDecision & control, simulation & learning provide tools that will helpto design an intelligent grid
We must move beyond the traditional static competitive equilibrium
analysis
We must move beyond the traditional myopic goal of economicefficiency, taking into account externalities such as sustainabilityand reliability
The best architecture for tomorrows energy highway?This is the central open question.
37/40 Conclusions
Concluding Remarks
http://find/http://goback/7/30/2019 Competitive Equilibria for Stochastic Dynamic Markets
87/90
An Entropic Grid may emerge as a result of many of the proposed
Smart Grid initiativesDecision & control, simulation & learning provide tools that will helpto design an intelligent grid
We must move beyond the traditional static competitive equilibrium
analysis
We must move beyond the traditional myopic goal of economicefficiency, taking into account externalities such as sustainabilityand reliability
The best architecture for tomorrows energy highway?This is the central open question.
Its solution opens many exciting research challenges!
37/40 Conclusions
Thanks!
http://find/7/30/2019 Competitive Equilibria for Stochastic Dynamic Markets
88/90
Celebrating with Dutch Babies after finishing part of this work
38/40 References
Control Techniques
Pre-publication version for on-line viewing . Monograph available for purchase at your favorite retailerMore information available at http://www.cambridge.org/us/catalogue/catalogue.asp?isbn=9780521884419
http://find/https://netfiles.uiuc.edu/meyn/www/spm_files/book.htmlhttps://netfiles.uiuc.edu/meyn/www/spm_files/CTCN/CTCN.html7/30/2019 Competitive Equilibria for Stochastic Dynamic Markets
89/90
FOR
Complex Networks
Sean Meyn
References
39/40 References
References
https://netfiles.uiuc.edu/meyn/www/spm_files/book.htmlhttps://netfiles.uiuc.edu/meyn/www/spm_files/book.htmlhttps://netfiles.uiuc.edu/meyn/www/spm_files/CTCN/CTCN.htmlhttp://find/7/30/2019 Competitive Equilibria for Stochastic Dynamic Markets
90/90
G. Wang, M. Negrete-Pincetic, A. Kowli, E. Shafieepoorfard, S. Meyn, and U. Shanbhag.
Dynamic competitive equilibria in electricity markets. In A. Chakrabortty and M. Illic,editors, Control and Optimization Theory for Electric Smart Grids. Springer, 2011.
M. Negrete-Pincetic and S. Meyn. Intelligence by design for the entropic grid. In IEEE PES11: Power Energy Society General Meeting, 2011.
G. Wang, A. Kowli, M. Negrete-Pincetic, E. Shafieepoorfard, and S. Meyn.A control theorists perspective on dynamic competitive equilibria in electricity markets. In
Proc. 18th World Congress of the International Federation of Automatic Control (IFAC),Milano, Italy, 2011.
S. Meyn, M. Negrete-Pincetic, G. Wang, A. Kowli, and E. Shafieepoorfard. The value ofvolatile resources in electricity markets. In Proc. of the 10th IEEE Conf. on Dec. andControl, Atlanta, GA, 2010.
I.-K. Cho and S. P. Meyn. Efficiency and marginal cost pricing in dynamic competitivemarkets with friction. Theoretical Economics, 5(2):215239, 2010.
M. Chen, I.-K. Cho, and S. Meyn. Reliability by design in a distributed power transmissionnetwork. Automatica, 42:12671281, August 2006.
40/40
https://netfiles.uiuc.edu/meyn/www/spm_files/MarketsIFAC2011/MarketsIFAC2011.htmlhttps://netfiles.uiuc.edu/meyn/www/spm_files/MarketsIFAC2011/MarketsIFAC2011.htmlhttps://netfiles.uiuc.edu/meyn/www/spm_files/MarketsIFAC2011/MarketsIFAC2011.htmlhttps://netfiles.uiuc.edu/meyn/www/spm_files/wind2010/Wind10.htmlhttps://netfiles.uiuc.edu/meyn/www/spm_files/wind2010/Wind10.htmlhttps://netfiles.uiuc.edu/meyn/www/spm_files/Market06/Market06.htmlhttps://netfiles.uiuc.edu/meyn/www/spm_files/Market06/Market06.htmlhttps://netfiles.uiuc.edu/meyn/www/spm_files/Market06/Market06.htmlhttps://netfiles.uiuc.edu/meyn/www/spm_files/Market06/Market06.htmlhttps://netfiles.uiuc.edu/meyn/www/spm_files/wind2010/Wind10.htmlhttps://netfiles.uiuc.edu/meyn/www/spm_files/wind2010/Wind10.htmlhttps://netfiles.uiuc.edu/meyn/www/spm_files/MarketsIFAC2011/MarketsIFAC2011.htmlhttps://netfiles.uiuc.edu/meyn/www/spm_files/MarketsIFAC2011/MarketsIFAC2011.htmlhttps://netfiles.uiuc.edu/meyn/www/spm_files/MarketsIFAC2011/MarketsIFAC2011.htmlhttp://find/