A Cooperative Market-based Decision Guidance Approach for
Resilient Power Systems
Alexander Brodsky, Eric Osterweil and Roberto Levy
Computer Science Department, George Mason University, Fairfax, VA, 22030, U.S.A.
Keywords: Resilience, Optimization, Power Systems, Cooperative Markets, Privacy, Security.
Abstract: National and local economies are strongly dependent on stable power systems. While the problem of power
system resilience in the face of natural disasters and terrorist attacks has been extensively studied from the
systems engineering perspective, a major unsolved problem remains in the need for preventive solutions
against the collapse of power systems. These solutions must ensure the most economically efficient operation
of power systems, within the bounds of any remaining power capacity. Transferring power usage rights from
the lowest-loss to the highest-loss entities would result in significant reduction of the combined loss. The
existing power systems do not take this fact into account. To address this need, we envision a paradigm shift
toward three-step system for (1) a cooperation power market where power usage rights can be transferred
among participating entities, (2) decision guidance to recommend market asks and bids to each entity, and (3)
optimization that, given the market clearance, will recommend precise operational controls for each entity’s
microgrid. The key challenge to address is the design of this three-step market system that will guarantee
important properties including Pareto-optimality, individual rationality, and fairness, as well as privacy,
security, pseudo-anonymity and non-repudiation.
1 INTRODUCTION
National and local economies are strongly dependent
on power systems, which involve power generation,
transmission, distribution and, increasingly,
distributed renewable power sources such as
photovoltaic arrays and wind turbines, local micro-
turbine generation, and power storage. The power
systems are getting increasingly complex, due to the
shift toward distributed and multidirectional flow of
power and largely unpredictable supply of power
from renewable sources, which are not dispatchable
(Moslehi & Kumar 2010). Figure 1 depicts a
prototypical electric power system with renewable
sources and power storage (US Energy Information
Administration.)
Power systems are highly vulnerable to natural
disasters and terrorism, resulting in huge economic
losses, such as the 2003 northeast blackout estimated
at $6 billion (Rose et al. 2007) (Lassila et al. 2005).
Sectors that are highly affected include non-durable
and durable manufacturing, construction, food
processing, wholesale trade and business services to
name a few (Rose et al. 2007).
Figure 1: Distributed power system with storage
technologies (Source: U.S. Energy Information
Administration).
While outages may occur due to natural causes,
such as hurricanes, blizzards, wildfires or technical
failures, preparedness for major disasters due to
terrorism is paramount. Unlike natural disasters or
technical failures, which occur randomly, terrorist
attacks, especially conducted by more sophisticated
state-supported players, can be optimized to cause
maximum damage with the payoff of high economic
impact and instilling fear (Rose et al. 2007). It is
256
Brodsky, A., Osterweil, E. and Levy, R.
A Cooperative Market-based Decision Guidance Approach for Resilient Power Systems.
DOI: 10.5220/0010309802560263
In Proceedings of the 10th International Conference on Operations Research and Enterpr ise Systems (ICORES 2021), pages 256-263
ISBN: 978-989-758-485-5
Copyright
c
 2021 by SCITEPRESS – Science and Technology Publications, Lda. All rights reserved

therefore critical to solve the problem of resiliency
and response to low-probability but extremely high-
impact scenarios (Wang et al. 2016).
While the existing power systems are far from
resilient, the problem of resiliency has been
extensively studied and is well-understood from the
system engineering perspective (e.g., see Wang et al.
2016, Guikema et al. 2006, Liu & Singh 2011, Wang
et al. 2015, Mohagheghi & Yang 2011) including (1)
hardening and resilience investment, such as
vegetation management, undergrounding and
elevating substations; (2) corrective actions and
emergency response, such as emergency load
shedding, special protection systems and islanding
schemes, and (3) damage assessment and system and
load restoration, such as distributed generation,
microgrids, distribution automation, mobile
transformers and decentralized restoration strategies.
A major unsolved problem remains, however, in
the need for mitigation solutions that incrementally
protect against the collapse of power systems as their
capacities become degraded. These mitigation
solutions must ensure most efficient operation of
power systems, in terms of their economic impact,
within the bounds of any remaining power capacity,
whether it is 80%, 50%, or 20% in large distribution
areas. The key difficulty in this problem is lack of
expressiveness into account for the economic impact
on diverse affected businesses and communities,
some of which may sustain huge economic losses,
while others would only be marginally affected.
Estimates of losses per kilowatt-hour (KWH)
from electricity disruptions range from $1.5 to
$7.5/KWH and, according to a more recent estimate,
at $50/KWH for some sectors (Rose et al. 2007). This
means, for example, that a business in these high-loss
sectors consuming 40 MW of power and losing half
of its power supply will be losing $1M per hour, while
a business at the lower end of the spectrum will be
losing “only” $0.03M per hour, with the combined
loss of $1.03M per hour.
It is easy to reduce this combined loss to “only”
$0.06M from $1.03M, by transferring 20 MW power
capacity from low-loss to high-loss business, and
compensating the low-loss business. This is over 94%
savings over the combined loss, but the power control
systems do not currently have the ability to take this
fact into account. Similarly, in the case of $1.5/kWh
vs. $7.5/kWh losses, we can save ⅔ of the combined
loss; in the case of $1.5/kWh vs. $4.5/kWh, we can
save ½ of the combined loss; and in the case of
$1.5/kWh vs. $3.0/kWh, we can save ⅓ of the
combined loss.
Optimally reducing the combined economic loss
through secure cooperation markets for power is
exactly the focus of this position paper. This is a
challenging problem, given the complexity of power
systems and diverse economic impacts to
participating (business, public or community)
entities. We believe, however, that this problem is
solvable, as discussed in the next sections.
2 PROPOSED SOLUTION
Power systems control, for every entity having a
microgrid (MG), is actuated for every time interval,
typically of 15, 30 or 60 minutes. Figure 2 depicts
prototypical microgrid components. The control deals
with how each power system resource/component is
operated, including: whether or not each power load
(e.g., for HVAC, lighting, data center) is activated
and at what level; whether each local generator is
activated and at what level of output; whether a power
storage device (e.g., high-capacity lithium battery) is
activated in charge or discharge mode and at what
level of power; and, the amount of power the entity
pulls from the grid (subject to contractual agreement
with a utility company), or possibly contributes to the
grid, when the power flow is reversed.
Figure 2: An Example of Microgrid Components.
We propose a three-step system for (1) a
cooperation power market where power usage rights
can be transferred among participating entities, (2)
decision guidance to recommend market asks and
bids to each entity, and (3) optimization that, given
the market clearance, will recommend precise
operational controls for each entity’s microgrid. The
main research challenge is the design of this three-
step market system that guarantees important
properties including Pareto-optimality, individual
rationality, fairness, as well as privacy, security,
pseudo-anonymity and non-repudiation.
A Cooperative Market-based Decision Guidance Approach for Resilient Power Systems
257

2.1 Three Step Cooperation Market
System
We envision the cooperation market be realized by
the three main steps:
● Cooperation market of power futures (or
options): This market runs before the beginning of
every time interval. Traded in this market are rights
to increase, or commitment to curtail, power
consumption upper bounds in time intervals 1,...,n for
each participating entity. The market clearance results
in (1) precise amounts of power in these
rights/commitments for each participating entity over
time intervals 1,...,n and (2) the amount of money
each entity receives from or gives to the market in lieu
of these rights/commitments. Of course, market
clearance must result in equilibrium of supply and
demand, for both power and money.
Before market runs, each participating entity submits
a combined (parametric) bid-ask, which we formulate
in the terminology of bids: agreeing to pay at most the
value v(𝑘𝑤

, …., 𝑘𝑤

) for the right to increase power
upper bounds by 𝑘𝑤

, …., 𝑘𝑤

) in time intervals
1,...,n. Note that, in this terminology, agreeing to pay
a negative amount, say -$1000, means receiving
$1000; and the right of power increase by a negative
amount, say -50 KW, means the commitment to
curtail power consumption upper bound by 50KW. A
bid-ask by a participating entity is a value function
v : [𝑚𝑖𝑛𝐾𝑊

, 𝑚𝑎𝑥𝐾𝑊

] x … x [𝑚𝑖𝑛𝐾𝑊

, 𝑚𝑎𝑥𝐾𝑊

] → R
(1)
where 𝑚𝑖𝑛𝐾𝑊

,..., 𝑚𝑖𝑛𝐾𝑊

are negative lower
bounds, 𝑚𝑎𝑥𝐾𝑊

, …, 𝑚𝑎𝑥𝐾𝑊

are positive upper
bounds, and the value v(𝑘𝑤

, …., 𝑘𝑤

) represents the
(maximum) amount of money the entity is ready to
pay for increasing the power consumption bounds by
(𝑘𝑤

, …., 𝑘𝑤

) in time intervals 1,...,n, relatively to
the current power upper bounds ( 𝑈𝐵

,..., 𝑈𝐵

)
allocated to the entity. Note again that 𝑘𝑤

< 0 means
that the entity will decrease the power upper bound
by 𝑘𝑤

in time interval i. And that v(𝑘𝑤

…., 𝑘𝑤

)
< 0 means that -v(𝑘𝑤

, …., 𝑘𝑤

) is the (minimum)
amount of money the entity is ready to receive for
(𝑘𝑤

,…., 𝑘𝑤

).
Given all bid-asks submitted in a market round,
market clearance results in precise
rights/commitments for power increase/curtailment
as well as payments made or received by participating
entities, as explained earlier.
● A decision guidance solution that, given a
description of all existing resources for an entity
(power loads and their equivalent values, local
generation, power storage, renewable sources, as well
as current power bounds), recommends the entity a
precise bid-ask to the market.
● A decision guidance solution that, given
market clearance, as well as the description of all
existing resources for an entity, performs value
optimization and recommends to the entity the precise
optimal operational parameters for each interval
1,...,n. The operational parameters include which
power loads are activated at what level (in KW) and
which are shed; which local generation resources are
activated at what level; which storage devices are
being used in charge or discharge mode and at what
level; etc.
2.2 Critical Properties of the Market
System
A major design challenge of the market system is to
assure some critical properties dealing with
optimality and fairness, which is easier to understand
in the framework of cooperative games. Consider a
cooperative game in which players who form a
coalition are entities that participate in the market.
Each entity (player) decides on bids/asks; then, the
market clears; finally, the entity decides on optimal
operation for n time intervals.
This optimal operation results in some value for
each participating entity, which is the total benefit of
running power loads (i.e., avoiding negative
economic impact) minus the total costs of operation,
plus (resp. minus) the money received from (resp.
given to) the market. If the entity does not participate
in the market, it can extract the value by optimizing
its resources within the available bounds of power
consumption. Let (𝑣

,..., 𝑣

) be the resulting values
for entities for the case when they do not cooperate
(i.e., do not participate in the market); and (𝑣


,..., 𝑣


)
be the resulting values for the entities if they do
cooperate, i.e., these are the values assigned to
players (entities) of the cooperative game (the market
system). A key research challenge is to design the
market system that will satisfy a number of important
properties of cooperative games, including the
following:
● Pareto-optimality (also called efficiency): it is
impossible to improve the resulting value 𝑣


of one
entity without sacrificing the value 𝑣


of at least one
other entity (i j). It is not difficult to show that this
property is equivalent to having operational
parameters of all entities that maximize the combined
benefit
∑
𝑣



. This is as though there were a
centralized authority that would jointly optimize all
entities and enforce the resulting operational
decisions across the board. This may be impractical,
ICORES 2021 - 10th International Conference on Operations Research and Enterprise Systems
258

particularly because business and communal entities
have very diverse power system and economic impact
characteristics, and, furthermore, may not want to
share this detailed (and sometimes confidential)
information with others or the centralized authority.
A major challenge that we need to overcome in the
envisioned market-based solution is the design of the
market system in which participating entities only
share their bid-asks yet the resulting operation of
power systems is equivalent to the result of joint
operational optimization (without actually doing it).
● Individual Rationality: 𝑣


𝑣

for every
entity i = 1,...,k. In other words, every entity can only
improve its value by participating in the game
(market), because otherwise it would not.
● Fairness: This property deals with how the
cooperation benefit
𝛥=
∑
𝑣




-
∑
𝑣



=
∑
𝑣


 𝑣




=
∑
𝛥



(2)
is distributed among the players (entities). Note
that, for each entity i, its cooperation benefit 𝛥

= (𝑣


-
𝑣

) must be non-negative to satisfy the individual
rationality property. An important question that needs
to be answered is how to define the most appropriate
definition of fairness in the context of power system
resilience and disaster response aimed at minimizing
losses. We believe that some proposed notions of
fairness, such as symmetry in cooperative games, may
not be appropriate, as we explain further in the
Technical Approach section. Intuitively, we would
like to have the notion of fairness by which low-loss
entities are compensated, possibly with some
premium, for their loss and agreement to further
curtail power consumption, to maximize loss
reduction of high-loss entities. This is as opposed to
the situation when low-loss entities would use a
disaster as an opportunity to make very high profits
(as opposed to cutting losses) due to them having a
valuable resource of power rights during a huge
shortage of power supply.
Privacy, security, confidentiality, pseudo-
anonymity, and non-repudiation: Transparency into
how the fair-market-price is computed and set will
build confidence in the fairness of the ecosystem, but
the underlying bid-asks expose some aspects of
market participants’ financial interests and
disposition. The privacy and security of this data
must be enforced by the implementing architecture.
The necessary properties of such an architecture will
be a model which will publish (to a subset of semi-
trusted audit/regulatory entities) and verify the
existence of any and all bid-asks that have been made,
pseudo-anonymity that will protect the privacy of the
identity of market participants associated with ask
and bid information, and provide non-repudiation (so
that once pseudo-anonymous bids and asks are
published and used to calculate the fair-market-value
of power, the submitting market participant cannot
disavow their existence).
3 TECHNICAL APPROACH
3.1 Design of Three-step Market
System
As described in the prior section, we need to perform
three tasks: (1) guiding entities on how to generate
their bid-asks; (2) market clearance given bid-asks
from all participating entities; and, (3) after the
market clears, optimization of operational controls of
power system for each entity. We discuss directions
for the solution, starting with market clearance.
3.1.1 Market Clearance
Assume that the bid-asks submitted to the market are
value functions 𝑣

,..., 𝑣

, where
𝑣

( 𝑘𝑤


,..., 𝑘𝑤


) represents the (maximum)
amount of money the entity i is ready to pay for
increasing the power consumption bounds by
(𝑘𝑤


,...,𝑘𝑤


) in time intervals 1,...,n, relatively to
the power upper bounds (𝑈𝐵


,...,𝑈𝐵


) currently
allocated to entity i. The first step is to determine
optimal flows of power
( 𝑘𝑤


,..., 𝑘𝑤


) for each entity i = 1,...,k by
maximizing the combined value of all entities
𝑣

𝑘𝑤


,...,𝑘𝑤





(3)
subject to:
● the (negative) lower bounds and (positive)
upper bounds on power transfer for every entity i =
1,...,k and every time interval j = 1,...,n, and
● power equilibrium (balancing) constraints, for
each time interval j =1,...,n.
This optimization results in power flow values
(𝑘𝑤

∗
,...,𝑘𝑤

∗
) for each entity i, as well as the
associated entity value 𝑣

∗
= 𝑣

( 𝑘𝑤

∗
𝑑,..., 𝑘𝑤

∗
).
This gives the market clearance for power flows. We
also need to determine the payment made or received
by each entity i = 1,...,k. To do that, consider the
optimal combined value V’ of all entities achieved by
cooperation: V’ =
∑
𝑣

∗

Also consider the
combined value V of all entities without cooperation:
V=
∑
𝑣



, where 𝑣

is the result of stand-alone value
A Cooperative Market-based Decision Guidance Approach for Resilient Power Systems
259

maximization for entity i, subject to its power usage
upper bounds. The difference 𝛥= V’ - V is the
cooperation benefit, i.e., increase in the combined
value due to cooperation. Depending on the fairness
criteria to be used, the cooperation benefit 𝛥 will be
distributed among the entities: 𝛥= 𝛥

+.... +𝛥

. Thus
the new value 𝑣


of each entity i must be the old value
(without cooperation) 𝑣

plus the cooperation benefit
𝛥

: 𝑣


= 𝑣

+ 𝛥

. But the value for entity i from the
combined value optimization is 𝑣

∗
. Therefore, the
payments paid or received by each entity i must be
done to cover the difference between𝑣


and 𝑣

∗
, so
that the cooperation benefit for entity i will be exactly
𝛥

, in accordance with the fairness criteria. This
completes the payment part of the market clearance.
While these are general ideas, we envision a careful
design and formalization of the three-step
cooperation market, developing its clearance
algorithm and mathematically proving that it satisfies
the desirable properties of Pareto-optimality,
individual rationality, and fairness.
3.1.2 Guiding Entity’s Decision on Bid-ask
to Market
We envision the development of a composite model
for power microgrid and its components, which will
express the value function V for the entity’s
microgrid. The value V(𝑘𝑤

, …., 𝑘𝑤

, 𝑈𝐵

,..., 𝑈𝐵

,
x), where x is a vector of all microgrid operational
controls over time intervals 1,...,n, is the total value
achieved by microgrid operation, which is the benefit
of running all activated load (e.g., preventing loss)
minus all costs. The model will also include the
Boolean function C(𝑘𝑤

, …., 𝑘𝑤

, 𝑈𝐵

,..., 𝑈𝐵

, x)
which expresses microgrid operational feasibility
constraint in terms of 𝑘𝑤

,…., 𝑘𝑤

, x.
Recall that operational control involve which
loads are activated at what level, which local
generators are activated at what level, which batteries
are activated in charge and discharge mode and at
what level etc. Given this information, we need to
compute bid-ask to the market, which is a function
v: [𝑚𝑖𝑛𝐾𝑊

, 𝑚𝑎𝑥𝐾𝑊

] x … x [𝑚𝑖𝑛𝐾𝑊

, 𝑚𝑎𝑥𝐾𝑊

] → R
where the value v(𝑘𝑤

,…., 𝑘𝑤

) represents the
(maximum) amount of money the entity is ready to
pay for increasing the power consumption bounds by
(𝑘𝑤

, …., 𝑘𝑤

) in time intervals 1,...,n, relatively to
the current power upper bounds ( 𝑈𝐵

,..., 𝑈𝐵

)
allocated to the entity.
We need to generate (a representation of) function
v under the assumption that, given (𝑘𝑤

, …., 𝑘𝑤

)
added to the existing power upper bounds, the
microgrid will be optimally operated. In other words,
v(𝑘𝑤

, …., 𝑘𝑤

) = 𝑚𝑎𝑥

V(𝑘𝑤

, …., 𝑘𝑤

, 𝑈𝐵

,..., 𝑈𝐵

, x)
(4)
subject to
(1) upper bound constraints for total power
consumption for every time interval,
(2) power balance constraints for every time
interval and
(3) microgrid operational constraints C(𝑘𝑤

, ….,
𝑘𝑤

, 𝑈𝐵

,..., 𝑈𝐵

, x).
The challenge in computing a representation of
this function is that it may not have a closed analytical
form, which is needed if we want to use efficient
mathematical programming algorithms (e.g., branch
and bound for mixed integer linear programming) in
market clearance optimization. Thus, we may need to
resort to its approximation. Computing this
approximation efficiently yet accurately is a research
challenge that needs to be addressed.
3.1.3 Optimization of Microgrid
Operational Controls
Market clears with instantiated power usage right
increases (𝑘𝑤

∗
, …., 𝑘𝑤

∗
) for an entity’s microgrid
for time intervals 1,...,n. Microgrid optimization is
maximization of the operational value V(𝑘𝑤

∗
, ….,
𝑘𝑤

∗
, 𝑈𝐵

,..., 𝑈𝐵

, x) for x subject to operational
constraints C(𝑘𝑤

∗
, …., 𝑘𝑤

∗
, 𝑈𝐵

,..., 𝑈𝐵

, x), when
all power usage right increases are fixed as per market
clearance. The challenge here is being able to
mathematically model microgrid operational value
and constraints for diverse set of resources used in the
microgrid. Also, since this optimization needs to be
done before the start of every time interval, efficiency
of an optimization algorithm is critical. We discuss
these challenges in more detail in the next section.
3.2 Power System Modeling, Decision
Guidance, and Optimization
All three steps in the three-step market system require
decision optimization. Finding bid-ask to be
submitted to the market and finding operational
control of the microgrid require modeling and
optimization of the underlying power system, as
described earlier. The model of the power system is
quite involved, because it must capture, in addition to
general computation of benefits, costs and balancing
constraints, the precise models of power network
components. They may include various types of
utility contracts; diesel generators; power storage
including batteries, spinning wheels, and hydro-
storage; schedulable loads such as ice generation for
ICORES 2021 - 10th International Conference on Operations Research and Enterprise Systems
260

cooling, boiling water and charging electric cars in a
parking garage; solar panels and wind turbines. These
models are quite involved and diverse.
To be able to scale the development of diverse
microgrid model instances, we would like to create a
reusable extensible repository of component models,
so that specific microgrid models can be easily
composed based on model components, similar to how
it is done in simulation-based systems (Lambert et al.
2006). At the same time, we would like to get
efficiency of the best mathematical programming
algorithms, such as for Mixed Integer Linear
Programming (MILP), which significantly outperform
simulation-based optimization algorithms. To bridge
the gap, we envision to leverage some research ideas
from our prior work, the work on microgrid component
models (Altaleb & Brodsky 2013, Levy et al. 2016b,
Levy et al. 2016a, Ngan et al. 2014), as well Unity
Decision Guidance Management System (Unity
DGMS) (Nachawati et al. 2017, Brodsky & Luo 2015).
It allows modular simulation-like modeling,
automatically generates mathematical programming
models, and solves them using the best available
mathematical programming algorithms. We plan to use
mixed integer linear programming solvers as well as
gradient-based non-linear programming solvers on
power system optimization.
To support the three-step market system we
envision the development of a Decision Guidance
solution based on Unity DGMS. The decision
guidance solution will be based on formal modular,
extensible analytic performance model which
expresses metrics of interest and feasibility
constraints as a function of investment and operation
decision variables. Metrics of interest include benefit,
cost and overall value of power system operation over
a number of time intervals. Feasibility constraints
include capacity limitation of physical resources,
power flow equation, contractual terms, and power
demand. Decision variables include all power system
operational controls over the planning time intervals
such as (1) power flows in the network as a whole, (2)
specific controls for each physical network
component such as power generators, transmission
lines, distribution, power storage, and renewable
sources of energy, and (3) financial instruments such
as contracts with power providers.
3.3 Market Privacy, Security,
Confidentiality,
Pseudo-anonymity, and
Non-repudiation
Entities involved in the energy marketplace will need
to expect that the market value of energy will be
computed fairly. The fair-market-value must be
computed by evaluating what each participating
consumer is willing to spend, how much energy is
needed, and (under appropriate circumstances) how
much power a customer may be willing to provide to
the power grid (and at what price). The ability to
audit how this price is computed and set will build
confidence in the fairness of the ecosystem.
Nevertheless, these data elements expose some
aspects of entities’ financial interests and disposition,
and the privacy and security of these data must be
enforced by the implementing architecture. To
ideally accomplish this, the information used will
need to be publishable to a set of entities (who may
not necessarily be market participants, but may be
regulatory), so that the market’s fairness can be
inspected and regulated. However, because exposure
of this level of consumer-interest in pricing would be
considered private information, it may result in
gaming of the system, and many market participants
may not want it to be publicly discoverable and
attributable, a pseudo-anonymous approach that
provides non-repudiation is critical. Such a viable
architecture will need to provide the necessary
transparency that allows inspection into how the fair-
market-value of energy was arrived at by a
community of auditing/regulatory entities, while also
protecting the security and privacy of consumers so
that their private data and interests maintain a
sufficient level of privacy protection, and must offer
non-repudiation facilities so that entities can be held
accountable after committing to asks/ bids.
The necessary properties of such an architecture
will be to create a model which allows market
participants to “bid” on energy (as previously
described), to be able to create an “ask” to provide
energy to the grid (as previously described). These
properties must also allow audit and regulatory
entities to verify the existence of any and all bids that
have been made by the set of market participants, the
existence of any and all asks that have been made by
the set of market participants, to protect the privacy
of sensitive information (such as the identity of
clients that can be associated with ask and bid
information), and non-repudiation (so that once bids
and asks are published and used to calculate the fair-
market-value of power, the submitting client cannot
A Cooperative Market-based Decision Guidance Approach for Resilient Power Systems
261

disavow their existence). In this architecture, it is
envisioned that there will be a set of semi-trusted
entities (such as the utility provider, possibly a set of
entities to compute the fair-market-value, or a
regulatory entity, etc.). This structure should ensure
that the marketplace and fair-market-value
computations are transparent enough that these semi-
trusted entities have only enough information to
verify bids and asks at admission time against the
pseudo-anonymous authors before publishing them.
The pseudo-anonymity architecture will be the
focus of additional research. With the data from bids
and asks being critical to computing and providing
transparency into the determination of a fair-market
value, the ability to disseminate these data to a semi-
trusted set of entities, for the data to be immutable,
and for it to be transparent will be explored in the
context of distributed ledgers. Initial considerations
will be given to blockchain technologies such as
private Ethereum, private bitcoin, and more recent
approaches such as those described by private
DLedger (Nakamoto n.d.; Zheng et al. 2017). Each
bid and ask will be separately represented in the
distributed ledger and will uniquely, and pseudo-
anonymously, indicate the specific market participant
who placed it. These investigations will underscore
the need for semi-public and immutable data to
bolster marketplace confidence, while still providing
privacy and non-repudiation.
To provide authentication and integrity
protections to the system’s semi-trusted providers, we
envision using their public DNS domain names (such
as example.com) and the DNS-based Authentication
of Named Entities (DANE) to build a reduced attack
surface security model (Osterweil et al. 2014). This
will allow protections to be managed by the power
utility, and will allow certificate’s to be issued to each
market participant. Each of these certificates will act
as a trust anchor for that entity and will correspond a
private key that will only be known to the market
participant whom the signing certificate was issued
to. These private keys will be used to create
attestations to revolving End Entity (EE) certificates,
which are created for every time interval of bids and
asks. These EE certificates will be used to create
digital signatures over each bid and ask, and that
signature will accompany its corresponding bid or ask
in the distributed ledger (with no other identifying
information). This will allow inspection of the data
in the ledger (by the subset of entities who are semi-
trusted), pseudo-anonymity of the market participants
(without the time interval-specific revolving EE
certificate, signatures do not identify the author’s
identity), and non-repudiation of each element (given
the EE certificate, each bid or ask can be verified and
attributed).
As an example, consider that a utility provider
issues subordinate signing certificates to k market
participants (𝑐

...𝑐

). At the beginning of each time
interval 𝑖, each client 𝑐

will create a new EE cert
(𝑐


). A bid 𝑏


and ask 𝑎


may be entered into a
distributed ledger with accompanying signatures
from that clients EE certificate from that epoch:
𝑏


,𝑠𝑖𝑔𝑛𝑎𝑡𝑢𝑟𝑒𝑏



Because 𝑐


is not publicly published, it is not
independently attributable. However, if a
compulsory audit is called for, each bid and ask can
be verified by having its corresponding EE certificate
disclosed, along with its pkcs7 signature chain to the
utility provider’s root certificate.
One key success criterion for the proposed
technology is scalability of optimization algorithms
for market clearance and microgrid operational
controls. This will be verified through a carefully
conducted experimental evaluation on a realistic
cooperation scenario to make an initial assessment on
the magnitude of economic losses that can be saved
during disaster response, as well as scalability of
optimization algorithms to deal with realistic size of
power systems in near real time.
Also, success criteria include the ability to
mathematically prove the desirable properties of
Pareto-optimality, individual rationality, and fairness
for the proposed three-step market system.
4 CONCLUSIONS
In a major shortage scenario, transferring power
usage rights from lowest-loss to highest-loss entities
has the potential of significantly reducing combined
loss and improving overall system resilience.
Unfortunately, the existing power systems do not take
this fact into account. With this drastic unutilized
reduction in combined losses, it is clear that power
systems need a paradigm shift, which we described in
this position paper. The market-based solution could
lead to a significant improvement of the resilience of
power cyber-physical systems.
REFERENCES
Altaleb, H. & Brodsky, A., 2013. A Primary Market for
Optimizing Power Peak-Load Demand Limits.
International Journal of Decision Support System
Technology, 5(2), pp.21–32. Available at:
ICORES 2021 - 10th International Conference on Operations Research and Enterprise Systems
262

http://dx.doi.org/10.4018/jdsst.2013040102.
Brodsky, A. & Luo, J., 2015. Decision Guidance
Analytics Language (DGAL) - Toward Reusable
Knowledge Base Centric Modeling. Proceedings of the
17th International Conference on Enterprise
Information Systems. Available at:
http://dx.doi.org/10.5220/0005349600670078.
Guikema, S.D., Davidson, R.A. & Liu, H., 2006. Statistical
Models of the Effects of Tree Trimming on Power
System Outages. IEEE Transactions on Power
Delivery, 21(3), pp.1549–1557. Available at:
http://dx.doi.org/10.1109/tpwrd.2005.860238.
Moslehi, K. & Kumar R., 2010, "A reliability perspective
of the smart grid." IEEE transactions on smart grid 1,
no. 1: 57-64.
Lambert T., Gilman P., Lilienthal P., 2006. Micropower
System Modeling with HOMER, Integration of
Alternative Sources of Energy, Farret, F, Godoy
Simões, John Wiley and Sons.
Lassila, J., Honkapuro, S. & Partanen, J., Economic
analysis of outage costs parameters and their
implications on investment decisions. IEEE Power
Engineering Society General Meeting, 2005. Available
at: http://dx.doi.org/10.1109/pes.2005.1489220.
Levy, R., Brodsky, A. & Luo, J., 2016a. Decision Guidance
Approach to Power Network Analysis and
Optimization. Proceedings of the 18th International
Conference on Enterprise Information Systems.
Available at:
http://dx.doi.org/10.5220/0005736401090117.
Levy, R., Brodsky, A. & Luo, J., 2016b. Decision guidance
framework to support operations and analysis of a
hybrid renewable energy system. Journal of
Management Analytics, 3(4), pp.285–304. Available at:
http://dx.doi.org/10.1080/23270012.2016.1229140.
Liu, Y. & Singh, C., 2011. A Methodology for Evaluation
of Hurricane Impact on Composite Power System
Reliability. IEEE Transactions on Power Systems,
26(1), pp.145–152. Available at:
http://dx.doi.org/10.1109/tpwrs.2010.2050219.
Mohagheghi, S. & Yang, F., 2011. Applications of
microgrids in distribution system service restoration.
ISGT 2011. Available at:
http://dx.doi.org/10.1109/isgt.2011.5759139.
Nachawati, M.O., Brodsky, A. & Luo, J., 2017. Unity
Decision Guidance Management System: Analytics
Engine and Reusable Model Repository. Proceedings
of the 19th International Conference on Enterprise
Information Systems. Available at:
http://dx.doi.org/10.5220/0006338703120323.
Nakamoto, N., Centralised Bitcoin: A Secure and High
Performance Electronic Cash System. SSRN Electronic
Journal. Available at:
http://dx.doi.org/10.2139/ssrn.3065723.
Ngan, C.-K. et al., 2014. Optimizing Power, Heating, and
Cooling Capacity on a Decision-Guided Energy
Investment Framework. Enterprise Information
Systems, pp.154–173. Available at:
http://dx.doi.org/10.1007/978-3-319-09492-2_10.
Osterweil, E., McPherson D & Zhang L, 2014. "The shape
and size of threats: Defining a networked system's
attack surface." In 2014 IEEE 22nd International
Conference on Network Protocols, pp. 636-641. IEEE.
Rose, A., Oladosu, G. & Liao, S.-Y., 2007. Business
interruption impacts of a terrorist attack on the electric
power system of Los Angeles: customer resilience to a
total blackout. Risk analysis: an official publication of
the Society for Risk Analysis
, 27(3), pp.513–531.
Wang, Y. et al., 2016. Research on Resilience of Power
Systems Under Natural Disasters—A Review. IEEE
Transactions on Power Systems, 31(2), pp.1604–1613.
Wang, Z. et al., 2015. Coordinated Energy Management of
Networked Microgrids in Distribution Systems. IEEE
Transactions on Smart Grid, 6(1), pp.45–53. Available
at: http://dx.doi.org/10.1109/tsg.2014.2329846.
Zheng, Z. et al., 2017. An Overview of Blockchain
Technology: Architecture, Consensus, and Future
Trends. 2017 IEEE International Congress on Big Data
(BigData Congress). Available at:
http://dx.doi.org/10.1109/bigdatacongress.2017.85.
A Cooperative Market-based Decision Guidance Approach for Resilient Power Systems
263