Modified Bat Algorithm based Approach for Optimization and

Allocation of Wheeling Prices under Open Access Power Market

Environment

KEYWORDS:, OPF-mile method, Simultaneous Bilateral Transactions, Multilateral

Transaction, Zero Counter Flow, Modified Bat Algorithm, Sellers and Buyers.

Abstract: This paper presents a modified Bat algorithm (MBA) based approach for optimization
of wheeling prices under open access environment of competitive electric power market. Usage
based allocation of optimized wheeling prices has been achieved by employing generalized and
topological distribution factors based tracing algorithms. Thus optimized wheeling prices has
been allocated across each of the users (suppliers and buyers both) of wheeling facilities. Used
wheeling capacity based OPF-mile pricing methodology has been proposed for providing correct
economic signals in the market and to motivate for the efficient use of wheeling facilities.
Absolute, zero counter flow (ZCF) and reverse variants have been considered to provide
diversity in wheeling pricing. On one hand wheeling pricing based on these variants establishes
the surety of maximum recovery of fixed wheeling cost. On the other hand it provides credits to
those network users, which are causing counter flows and thus helping in releasing congestion.
The proposed work has been carried out for several bilateral and multilateral transactions along
with pool power paradigm. It has been demonstrated on IEEE 30-bus system and an Indian
Utility 62-bus system. The optimized wheeling prices obtained with proposed MBA-OPF have
been compared with conventional OPF results and with a common metaheuristic technique i.e.
GA based OPF results and found satisfactory.

KEYWORDS: Generalized Distribution factors, Topological Distribution factors, AF-mile,

Hybrid Power Market, Simultaneous Bilateral Trading, Multilateral Trading, Zero Counter
Flow, Modified Bat algorithm.

1. Introduction: Pricing of wheeling services has now become a very important issue in recent
unbundled business environment of electric supply industry. It has a significant impact on future
expansion of existing wheeling facility and to make right financial settlements for all market
participants (suppliers and buyers). The process of converting investment cost of wheeling
infrastructures into wheeling charges is termed as wheeling pricing paradigm. This paradigm
includes three cost based pricing techniques, which are embedded cost based pricing,
incremental cost based pricing and composite embedded/incremental cost based pricing
techniques [ref1]. Out of these techniques, embedded cost based pricing has gained much
popularity as it was able to recover sufficient financial revenue including embedded capital
investment, operation and maintenance cost of existing wheeling facilities [ref2]. Several
methods based on embedded cost have been found in the literature. These are Rolled-In-
Embedded methods, Boundary-flow method and Line-by-line methods.

Postage stamp method and Contract path method come into the category of Rolled-In-
Embedded methods. These methods provide only average system cost and do not reflect correct
economic signals, because they do not perform power flow execution in determining pricing.
These methods have been criticized as having no obvious economic background and ignoring the
actual system operation condition [ref3]. Other two embedded cost methods which require actual
power flow execution are Boundary-flow method and Line-by-line (original MW-mile) method.
MW-mile original method is supposed to be the first method proposed to recover fixed wheeling
cost, based on the actual use of wheeling network [ref4].

In the competitive electrical power market, suppliers and buyers are required to establish
the appropriate wheeling prices in order to make efficient use of wheeling facilities and to make
better profit margins through wheeling. Since charging a wheeling transaction mainly depends
upon overall charge of availed wheeling facilities, a significant research has focused on usage-
based cost allocation [ref5, ref6]. Usage based evaluation of wheeling facility cost is very
difficult to obtain, due to non linear nature of power flow equations. However some methods are
found in the literature which provide possibly acceptable solutions for determining contribution
to the network flows from individual users (suppliers and buyers). These methods are
Distribution factors based method [ref7], Sensitivity indices based method [ref8]. Power flow
decomposition based method [ref9] and Tracing algorithms [ref10, ref11, ref12].

For monitoring both real and reactive power flow to determine actual usage based
wheeling prices, MVA-mile method was suggested by researchers [ref13]. An improved version
of MW-mile and MVA-mile was proposed, in which instead of multiplying length of line, the
average revenue requirement per hour of the line was multiplied to obtain transaction cost per
hour [ref14]. A modified Amp-mile and MVA utility factor based embedded cost allocation
method was presented in [Ref 15][Gaurav Jain 2019].

Usage based allocation of wheeling prices involves charging of used transmission

capacity and un-used transmission capacity[ref16, ref6]. In the un-used transmission capacity
method, users have to pay for whole transmission capacity whether it is being utilized or not. On
the other hand, used capacity methods charge users only for actual line flows they cause, not for
whole transmission capacity. So the later one provides inspiration for efficient use of available
wheeling facility even though it cannot guarantee the full recovery of fixed wheeling cost.

A reduction in wheeling cost was achieved by reducing power losses through optimal
placement of recent FACTS device i.e. IPFC Controller [ref17]. A DC-OPF based MW-mile
method along with various tracing techniques to allocate wheeling prices among market
participants has been presented in [ref 16]. The purpose of employing OPF program is to
determine optimized flows in a wheeling network as a result of particular wheeling transaction so
that wheeling prices can also be optimized. Though the conventional optimization techniques
like Interior Point method, Gradient method, Linear programing, Quadratic programming can be
applied to solve optimal power flow problem [ref18,19,20], these techniques have their own
drawbacks. Some of the drawbacks are improper or local convergence, sometimes no
convergence and requiring large number of iterations; resulting large computational time. On the
other hand, metaheuristic techniques have gained great research attention as an alternative to the
conventional optimization techniques being tolerant of non-convex, non-differentiable and
reasonable practical computational time [ref21][ref LS mam first paper].

During recent years several metaheuristic techniques have been developed which are up
to some extent can be classified as nature inspired algorithms [22][Meng 2015] . These
algorithms have drawn some abstraction from nature, including animals, plants, natural
phenomena or law. Some examples of these algorithms are Genetic algorithm (GA) [ref23][ snp
me se], Particle Swarm Optimization (PSO) [ref24, ref25], Bacterial Foraging Optimization
(BFO)[ref26], Bat algorithm (BA) [ref27] etc. GA based power flow tracing approach was
proposed for wheeling pricing based on MVA-km method in [ref28]. GAOPF based MW-mile
and MVA-mile methods for allocating wheeling prices across various users of the system, have
been discussed in [ref29]. In the present paper a modified Bat algorithm (MBA) [ref22, ref30]
has been proposed for optimizing wheeling prices. In the proposed modified bat algorithm,
Doppler effect has also been incorporated, means each virtual bat can now adaptively
compensate for the Doppler effect in echoes. In original BA, virtual bats search their food in one
single habitat with mechanical behavior. However in the modified BA, virtual bats are proposed
with diversified foraging habitats with quantum behavior by the virtue of which the bat can
occupy any position in the whole search space with certain probability [ref22]. The proposed
approach has also been compared with conventional OPF method and GA-OPF method in regard
of computational time and degree of optimization and found satisfactory.

Two tracing algorithms i.e. Rudnick tracing and Bialek Tracing have been implemented
in this work to allocate optimized wheeling prices across each user of wheeling network. Thus
this work presents the optimized wheeling prices for the given wheeling facility across each
supplier and buyer in a pool power market paradigm and for various feasible bilateral and
multilateral transactions. In Rudnick tracing generalized distribution factors represents the
impact of particular generation [ref 31] [ref Ng] or load [ref7 rudnick] on a specific line flow,
which could be negative. Thus this tracing can account the counter flows in wheeling pricing
while Bialek tracing is based on proportional sharing basis in which topological distribution
factors shows only the share of a particular generation in the total line flow. This share can not
be negative, it is a quantitative term [ref 10][ref bialek97]. Hence Bialek tracing cannot
incorporate counter flows in wheeling pricing.

The used wheeling capacity based OPF-mile methodology for optimizing wheeling prices
has been proposed with three variants of flows viz absolute, ZCF and reverse. Absolute variant
of wheeling pricing is beneficial for wheeling network owner authority, ZCF variant is beneficial
for users (suppliers and buyers both) and reverse variant introduces benefit for those users which
are creating counter flows. The present approach of optimization and allocation of wheeling
prices across each supplier and buyer of the network has been demonstrated on IEEE 30-bus
system [ref32] and Indian utility 62-bus system [ref33].

.
 The present paper proposes a MBA based OPF technique for optimization of wheeling
prices considering charges for counter flows in three different variations. Allocation of these
prices among market participants has been accomplished through two tracing techniques and
results are demonstrated on IEEE 30-bus system [ref30] and Indian utility 62-bus system
[ref31]. The proposed approach has also been compared with conventional OPF method and GA-
OPF method in regard of computational time and degree of optimization and found satisfactory.

OLD

Due to worldwide restructuring, electricity wheeling service pricing is highly prioritized and
important task for all types of electrical power industries. It has a significant impact to make
right financial settlements for all market participants and expansion of transmission system. In an
emulous electrical power market, transmission services are mainly classified into pool, bilateral
and multilateral trading [1].

In a competitive electrical power market price, stability and certainty are necessary to perform
long term planning and rationally social optimal investment. However, pool power market
provides a better resource usage but due to centralized market place, price and revenue certainty
and easy implementation of support for new and renewable resources are not possible. Therefore,
pool with bilateral and multilateral power markets has positive effects on the functioning of
electricity markets and benefits all users. The California, Nordic, New Zealand and New York
models etc. fall into this category [1, 2].

All types of power markets are subject to economies of scale. So, on the basis of tariff, global
conditions and revenue requirement wheeling pricing methods can be categorized into
embedded, marginal and incremental cost methods [3, 4]. To find sufficient financial revenue,
embedded cost methods would always be preferred because of no assurance of sufficient gain in
marginal and incremental cost methods [5, 6]. Embedded cost methods are mainly classified into
two categories which are based on power flow execution. The first type sways around embedded
cost methods which do not require power flow execution [7]. The second method, which depends
on power flow execution i.e. boundary flow method and line by line method, gives more accurate
and actual results in grid/transmission sector [8, 9].

In line by line method, apparent flow mile (AF-mile) methodology is more prominent as
compared to MW mile methodology because it considers both real and reactive power [10, 11,
12]. Line by line method is mainly classified into two types- full capacity based method and used
capacity based methods. In full capacity based method, users pay not only for the actual line but
also for unused capacity while in used capacity based method, users pay only for the actual line
flows they cause [13, 14]. Due to the presence of willing buyer and seller in a competitive power
market, efficient or actual use is mostly liked to happen. Three options with AF mile method
have been presented in this paper for efficient allocation of wheeling prices i.e. used absolute,
used Zero Counter Flow (used ZCF) and used reverse methods.

Generally in bilateral and multilateral power markets, wheeling prices allocation can be
performed for a specific transaction. But due to bundled grid/transmission system, single
transaction response appears across all users in a transmission system. Therefore, it is important
to allocate wheeling prices across all users for a specific transaction. If wheeling prices are
allocated across all users in a competitive power market then power tracing techniques become
an important tool for price allocation. In electrical power market ,various types of tracing
techniques are available such as Rudnick tracing, Bialek tracing, Kirschen’s tracing, minimum
power distance tracing ,etc. but all power tracings are not equally applicable everywhere. Also,
there are various limitations with different tracing methods.

Rudnick tracing or generalized distribution factors based method is traditionally used for power
system security and contingency analysis, but according to the established literature, Rudnick
tracing with AF-mile methodology is capable to resolve maximum transaction related power
flow [5, 15]. Besides this, Bialek tracing is based on proportional sharing principle and to
determine topological distribution factors for users with transaction related power flow [16, 17].
In Kirschen’s tracing, network elements are organized in homogeneous groups and are
applicable for power system security constraints [18, 19, 20]. Whereas minimum power distance
tracing method minimizes the total power distance of line for assessment of wheeling prices [21].

In this paper, assessment and allocation of wheeling prices across each user have been
implemented in simultaneous bilateral and multilateral power market where one or many power
markets may be a part of hybrid power market. This paper investigates limitations and feasibility
of two well known tracing techniques- Rudnick tracing and Bialek tracing in a competitive
power market scenario, because both are used to resolve transaction related power flow. The
effect of above tracing techniques has been shown with actual and efficient results by using three
used capacity based methods i.e. used absolute AF-mile, used ZCF AF-mile and used-reverse
AF-mile.

The above approach of wheeling pricing with feasibility and limitations of tracing techniques has
been investigated on IEEE-30 bus system [22] and Indian utility 62 bus test system [23].

2. Problem Formulation

2.1 Optimal Power Flow: - Actual and optimized power flow is an integral factor of embedded
cost line by line method for assessment and allocation of wheeling prices. Power flow using
embedded cost method gives more realistic and efficient results in a competitive power market
[8, 9]. Several conventional methods are available in literature to solve optimal power flow like
Gradient method, Hessian method, Linear programming method, Quadratic method and interior
point method [24, 25, 26]. But to calculate more probabilistic results, intelligent system based
methods have been preferred in place of conventional methods. In this paper, one of the
prominent modified Bat algorithm based optimal power flow method has to be suggested, which
can stuck in local minima and make the convergence faster.

2.1.1 Modified Bat algorithm based OPF: - Different types of heuristic and Meta -heuristic
techniques are in circulation to calculate the optimal power flow. The modified bat algorithm
based technique suggested in the paper is a modification of the original bat algorithm. In which
bats have self adaptive compensation of Doppler Effect in echoes. This compensation rate
depends upon the species of bats. Some species have partial compensation while other can
exhibit close to full compensation [27, 28].

Also the foraging habitats of bats vary the species. Bats can adjust their echolocation behavior

under different habitats in which they forage. Some bats species forage in a wide range of
habitats, while other may forage mainly in water or forest habitats.

In the Modified Bat algorithm, Doppler Effect and diversified foraging habitats have been seen
incorporated to present the algorithm from being stuck in local minima and make the
convergence faster.

STEP 1:- For iteration t=0, initialize the population and related parameters.

STEP 2:- Evaluate the objective function and determine fitness of each individual.

STEP 3:- For iteration in the maximum count (i.e. for t < m), invoke habitat selection, by

comparing probability of habitat selection (i.e. P) with a uniform random no. rand1 [0, 1].

STEP 4:- Implement the local search by considering the loudness of surrounding environment
apart from loudness produced by the bat itself.

STEP 5:- evaluate the fitness of each individual.

STEP 6:- Update solutions, loudness and pulse emission rate globally.

STEP 7:- Rank the solution and find the current best .
STEP 8:- If does not improve in G time steps, means when all the bats do not find a better
prey than the previous one in several time steps G, this is the situation of being trapped in local
optima. Then the bats would fly to another place to search for prey.

STEP 9:- increase the iteration count by 1.

STEP 10:- if termination criteria is met, finish the algorithm otherwise go to step 3.

STEP 11:- Obtain the individual the best objective function value in the population.

2.1.2 Modeling of Transactions: Before assessment and allocation of wheeling pricing, it is

necessary to determine feasible transactions. The conceptual modeling of wheeling transaction is
that sellers and buyers encourage the trading between them without disrupting the transmission
constraints [29]. Mathematically, each bilateral exchange of power either proposed, scheduled or
actual one between a seller bus-i and buyer bus-j satisfies the following power balance
relationship.

(1)

Fig.1 Modeling of Transactions with MBA based OPF

In case of a multilateral exchange, power is injected in different buses and taken out from
different buses simultaneously. The summation of power injected in different buses-i is equal to
the summation of load powers taken at various buses- j.

(2)

th th
Where, is seller apparent flow at bus and is buyer apparent flow at bus.

2.2 Tracing of Wheeling Prices: - In deregulated power market, production (generation) and
demand (load) cannot be constant ever. When production and demand have been changed or
shifted, conventional load flow analysis could not trace the electricity [2]. To solve this problem,
various types of methods are used to calculate line flows and allocation of wheeling prices
namely Rudnick method, Bialek method, Kirchen method, minimum power distance method, etc.
All these methods are called tracing methods. Most of the tracing methods in power market have
been applied for security analysis but in transaction (or market) related power flows, all methods
cannot be equally applicable. In this paper, two tracing methods have been applied to allocated
wheeling prices with transaction related power flows i.e. Rudnick tracing and Bialek tracing
methods. Mostly wheeling prices have to be calculated from one bus to another bus (line by
line). But in hybrid power market, allocation of wheeling prices across each user with separate or
simultaneous transaction can play a vital role to analyze the overall system of cost of market.
This paper investigates the approach of wheeling pricing assessment in a competitive power
market (bilateral and multilateral) across each user, which would be the non-separable part of
hybrid power market and not yet found in the literature. After the assessment of wheeling prices
across each user in bilateral and multilateral power market, overall system prices can be analyzed
by employing pool and bilateral/multilateral constraints as per the requirement. With this,
allocation of individual user price will also be possible at any transacted power flow.

Rudnick tracing method is based on generalized distribution factors while Bialek tracing method
is based on topological distribution factors. To establish the significance and limitations of both
the methods, three individual wheeling pricing methods are employed. Through these methods,
efficient wheeling prices have been allocated across each user in competitive as well as hybrid
power market.
2.2 Flow based wheeling pricing: - AF-mile method has been proposed for the recovery of
MW-mile method, in which reactive power flow is not considered [29]. AF-mile method can
send more realistic and economical signal to all the users. The price allocation of kth wheeling
facility of power flow based line by line using AF-mile method is formulated as:

(3)

where WP is the wheeling prices allocation of kth wheeling facility in k$, ck is the predetermined
embedded cost of kth wheeling facility for pu length , Lk line length in mile, is the
maximum capacity of line k and (AF)k is the apparent flow of kth wheeling facility.

2.2.1 Variants of AF-mile Methods

Variant means exhibiting variety or diversity. In used AF-mile method, three types of variants
are possible.

2.2.1.1 Absolute Variants

The absolute variants of apparent flow mile methods are helpful to calculate prices on the actual
magnitude of power flow irrespective of its direction. The (AF)k is the flow by kth wheeling
facility and formulated as:

(4)

For apparent flow in any direction

2.2.1.2 Zero Counter Flow Variants

The dominant variants of apparent flow mile methods are helpful to calculate prices for only
positive power flow. This method is known as hybridization of absolute and reverse method. In
this, reverse flow of power has been considered as zero. This method is not easily accepted by
grid owner to calculate prices. The (AF)k is the flow by kth wheeling facility and formulated as:

(5)
For apparent flow in positive direction

(6)

For apparent flow in reverse direction

2.2.1.3 Reverse Variants

The reverse variants of apparent flow mile methods are helpful to calculate prices of the
magnitude and direction of the flow. The effect of reverse flow of power reduces congestion in
the lines. The (AF)k is the flow by kth wheeling facility and is formulated as:

(7)

For apparent flow in positive direction

(8)

For apparent flow in reverse direction

3. Research Methodology: Wheeling price assessment and price optimization is the most
important material to increase the efficiency of power transmission and reduce congestion.

In order to understand the meaning and functioning of the article well, Fig. 2 is given. Here in
Fig. 2 the resultant power flow is calculated by optimizing the fitness function by MBA based
OPF. Initially line data, bus data, cost function coefficients, line capacity and line cost are given
in the system as input to calculate the resultant power flow.

After receiving the optimized apparent flow, two most popular rudnick tracing and bialek tracing
are used here for price allocation.

The wheeling prices for all sellers and buyers have been calculated in the paper by generalized
generation distribution factor (GGDF) and generalized load distribution factor (GLDF) in
Rudnick tracing and with the help of upstream and downstream in Bialek tracing.

Here the final price allocation is done with used absolute, used ZCF and used reverse methods as
shown in figure 2.
 Fig.2 Schematic block diagram to demonstrate functionality

The flow chart is shown in Fig. 3 to understand the overall activity and suggested optimization
technique. In this, most commonly used cost function of each generating unit has assumed a
fitness function. This fitness function has optimized with provided input data in specific limits.
 Fig.3 Flow chart of overall methodology

4. Results and Discussion: Power flow optimization is required for actual flow based wheeling
price assessment and allocation. In this paper, two tracing techniques (Rudnick and Bialek) and
two power flow optimization techniques (GA based OPF and MBA based OPF) have been used
for the allocation of value across each user. This paper demonstrates the importance of the
proposed method of wheeling flow pricing and allocation by three types of variants (absolute,
ZCF and reverse) and two power flow optimization techniques. By making bilateral and
multilateral transactions with all the above variants, optimization techniques, and tracing
techniques, this paper attempts to propose a new approach to the hybrid competitive power
market. The IEEE 30-bus system and the Indian utility 62-bus system have been used in the
paper to test the above approach.

If first look at the IEEE 30-bus system, as per the data, total annualized cost of IEEE-30 bus
system is 8851.84k$. In the present power market 06 total seller buses and 21 buyer buses have
been used to allocate wheeling service pricing. Total numbers of wheeling tracks are 41 which
are helpful to wheel power in network. For calculation purpose, assume the seller’s contribution
as 30% while the buyer’s contribution is 70% of the total wheeling cost. In this way, contribution
of the seller is 2655.55 k$ and contribution of the buyer is 6196.29 k$, with respect to total fixed
cost of transmission. For assessment of market nature and evaluation of power market,
transactions are considered separately for simultaneous bilateral power market and then two
multilateral transactions TM1 and TM2 in power market. The details of transactions are as
follows:

Table 1: Feasible Simultaneous bilateral and multilateral Transactions for IEEE 30-bus
system
Feasible Simultaneous bilateral Transactions
S.No. Transaction From Bus To Bus Value of transaction
1 TB1 10 8 15MW
2 TB2 4 12 17MW
3 TB3 15 21 08MW
4 TB4 12 16 10MW
Total Simultaneous Bilateral Transactions TSB 50MW
Feasible Multilateral Transactions
S.No. Transaction From Bus Value of To Bus Value of
transaction transaction
1 TM1 08 25MW 13 20MW
06 45MW 19 25MW
18 30MW 22 30MW
11 10MW
14 15MW
Total Multilateral Transactions TM1 100MW 100MW
2 TM2 07 30MW 06 20MW
21 35MW 18 15MW
23 30MW
Total Multilateral Transactions TM2 65MW 65MW
Table 2: Feasible Simultaneous bilateral and multilateral Transactions for Indian utility
62-bus system
Feasible Simultaneous bilateral Transactions
S.No. Transaction From Bus To Bus Value of transaction
1 TB1 14 6 20MW
2 TB2 26 04 18MW
3 TB3 16 07 17MW
4 TB4 04 17 15MW
Total Simultaneous Bilateral Transactions TSB 70MW
Feasible Multilateral Transactions
S.No. Transaction From Bus Value of To Bus Value of
transaction transaction
1 TM1 05 20MW 08 25MW
11 25MW 17 10MW
21 55MW 13 35MW
28 20MW 31 20MW
15 20MW
17 10MW
Total Multilateral Transactions TM1 120MW 120MW
2 TM2 08 35MW 11 15MW
17 45MW 15 20MW
19 15MW
27 10MW
21 20MW
Total Multilateral Transactions TM2 80MW 80MW

After taking the above transactions and seeing them in table 3, it is found that GAOPF and
Rudnick tracing under the used absolute method in pool power market determine the total price
of 2476.92 K$ for all sellers. With this here, buyers get a price of 6788.08K$. The total price of
the seller and buyer (9265.00K$) exceeds the transmission cost determined in this method by
8851.84K$, which is in the interest of the transmission owner but not entirely suitable for
customers. Since both positive and negative flows are evaluated in the method used, it is
necessary for transmission owners and customers to evaluate the final value while
simultaneously taking care of the fixed transmission. This type of evaluation is necessary to
protect the interests of both transmission owners and customers. Competitive power To
demonstrate market thinking realistically, When we look at Table 3 after imposing available
bilateral transactions in Table 1, we find that the total value of all sellers is 2449.16K$ and all
The buyer has a total of 6742.42K$. So the wheeling price is slightly lower than the pool power
market due to the low power flow values in the simultaneous bilateral transmission. But this
value is also higher than fixed transmission cost, which does not fully protect the interests of the
customers.
In addition, by looking at the result of the multilateral transactions shown in Table 1 and Table 3,
we find that the values of TM1 and TM2 of all sellers are 2378.99K$ and 2401.05K$
respectively and that of all buyers are 6725.35K$ and 6740.00K$. All the results obtained
depend on the actual power flow and show the corresponding values of the power flow. Any
value found in the used absolute method, whether it is pool, simultaneous bilateral or multilateral
transactions, displays a value higher than the fixed transmission cost.

Table 3: Wheeling Prices Allocation in IEEE-30 Bus System based on Simultaneous bilateral and multilateral Power Market with Base Case using
Rudnick Tracing and GAOPF

Wheeling Pricing Methods ( wheeling prices in k$)

used abs used rev

used ZCF
TB1+TB used used TB1+TB2
used abs used abs TB1+TB2 used rev used rev
S. 2+TB3 used ZCF ZCF used +TB3
used abs TM1 TM2 +TB3 ggdf TM1 TM2
No Users ggdf ZCF TM1 TM2 reverse ggdf
ggdf multilate multilate simultane multilate multilate
. simultan ggdf multilate multilate ggdf simultane
ral ggdf ral ggdf ous ral ggdf ral ggdf
eous ral ggdf ral ggdf ous
bilateral
bilateral bilateral

1 S1 1548.66 1630.26 1606.40 1707.18 1494.26 1455.37 1446.37 1549.67 1439.35 1489.99 1525.30 1509.72
2 S2 318.19 295.29 280.92 293.66 283.29 242.73 232.77 245.50 250.00 227.08 224.18 218.61
3 S5 131.27 124.90 118.57 123.82 62.07 54.17 51.76 54.58 -2.53 -3.32 -3.49 -3.22
4 S8 178.34 144.23 129.11 86.76 86.36 64.21 63.95 39.26 0.50 -0.44 -0.60 -0.28
5 S11 153.33 120.14 118.00 80.96 89.48 64.62 57.88 44.25 29.81 22.44 22.77 14.61
6 S13 147.13 134.34 125.99 108.69 92.55 77.92 73.64 64.07 41.48 36.71 35.64 28.69
Total 2476.92 2449.16 2378.99 2401.05 2108.01 1959.02 1926.38 1997.33 1758.62 1772.46 1803.80 1768.13
7 B2 178.73 172.39 171.20 163.31 87.62 90.50 90.54 95.95 4.25 14.80 17.64 27.43
8 B3 35.81 35.03 34.87 34.33 13.54 13.92 13.90 14.88 -7.48 -5.99 -5.63 -4.44
9 B4 105.91 104.55 104.31 103.64 60.35 61.82 61.60 65.87 20.00 22.96 23.83 27.04
10 B5 1531.21 1524.53 1522.15 1520.79 1307.79 1326.17 1316.08 1394.51 1191.88 1190.80 1195.22 1223.90
11 B7 419.85 418.76 418.12 420.39 316.25 321.42 319.08 339.59 238.99 240.71 242.11 249.64
12 B8 621.56 625.77 626.33 638.14 475.67 485.12 482.01 516.11 369.33 369.40 370.87 380.12
13 B10 175.64 174.84 174.32 175.44 132.77 133.85 132.51 140.25 100.96 99.80 99.88 101.33
14 B12 324.97 321.32 320.20 320.14 203.49 205.07 203.16 214.91 99.34 100.91 101.72 105.62
15 B14 237.14 234.05 233.08 232.60 147.51 147.98 146.41 154.22 70.44 70.69 71.04 73.02
16 B15 296.70 292.95 291.77 291.46 211.47 211.90 209.59 220.62 143.98 142.27 142.39 144.41
17 B16 109.03 108.31 108.02 108.55 79.69 80.26 79.48 84.06 57.01 56.47 56.55 57.46
18 B17 295.53 293.76 292.84 294.33 243.17 244.35 241.74 255.09 210.87 206.90 206.66 208.26
19 B18 123.69 122.52 365.84 122.33 96.63 96.84 95.78 100.83 77.60 76.08 75.98 76.52
20 B19 370.21 367.14 122.11 366.90 319.29 319.75 316.08 332.48 294.60 287.53 286.79 287.63
21 B20 81.26 80.64 80.37 80.66 64.48 64.70 63.99 67.44 53.05 52.02 51.94 52.31
22 B21 664.53 659.33 657.00 659.11 541.84 543.29 537.20 565.70 463.91 454.02 453.20 455.67
23 B23 129.60 128.31 127.86 128.03 98.82 99.00 97.90 103.03 76.27 74.80 74.72 75.27
24 B24 345.87 342.71 341.50 342.81 290.99 291.32 287.99 303.15 260.07 253.98 253.37 254.26
25 B26 166.18 164.88 164.39 165.08 134.34 134.50 132.99 139.94 113.61 110.80 110.50 110.75
26 B29 104.39 103.68 103.40 103.95 77.43 77.72 76.89 81.11 56.94 55.85 55.78 56.19
27 B30 470.28 466.97 465.67 468.02 371.28 372.20 368.12 387.83 303.05 296.18 295.54 296.78
Total 6788.08 6742.42 6725.35 6740.00 5274.44 5321.69 5273.05 5577.58 4198.68 4170.97 4180.10 4259.15
If we look at the values obtained from the used ZCF method, then the achievable value pool of
the sellers, simultaneous bilateral and multilateral transactions TM1 and TM2 have 2108.01K$,
1959.02K$, 1926.38K$, 1997.33K$ respectively. The corresponding buyers' prices here are
obtained as 5274.44K $, 5321.69K $, 5273.05K $, 5577.58K $ respectively. All values of the
ZCF method used are based only on positive flow and are therefore not entirely appropriate from
the transmission owner's point of view because of the failure to recover fixed transmission costs.
If we look at the values obtained from used reverse method, then the achievable value pool of the
sellers, simultaneous bilateral and multilateral transactions TM1 and TM2 have 1758.62K$,
1772.46K$, 1803.80K$ respectively. The corresponding values of the buyers are obtained by
4198.68K$, 4170.97K$, 4180.10K$, 4259.15K$ respectively. In used reverse method, some of
the values on buses are obtained in wheeling price minus. The reason for this is that due to the
credit of negative flow to the customer in this method, sometimes the credit amount exceeds the
price of positive flow and the customer gets extra incentives. This method does not achieve the
status of practical method due to it not being suitable at all from the perspective of transmission
owner.

Fig.4 Allocation of Wheeling Prices for Sellers on IEEE-30 Bus System using Rudnick Tracing
with GAOPF
All the above values are values derived from Rudnik tracing which are able to calculate the price
obtained from negative flow. By comparing all the above values obtained with GAOPF based
method to suggested method, the results obtained can be reviewed for betterment and feasibility.
The values obtained for all sellers and buyers are compared and demonstrated by bars in Figures
4 and 5.

Fig.5 Allocation of Wheeling Prices for Buyers on IEEE-30 Bus System using Rudnick Tracing
with GAOPF

The above figures include graphs between users and the received cost, which ensures comparison
of all the received methods on one platform. In the next step if we compare the GAOPF based
Wheeling Pricing Approach to the new approach MBA based Wheeling Pricing Approach
suggested in the paper. Here we find that in the MBA absolute wheeling pricing approach in the
used absolute method, the values of pool, simultaneous bilateral and multilateral TM1 and TM2
are obtained for all sellers by 2403.69K$, 2386.71K$, 2332.28K$ and 2393.38K$ respectively.
Similarly, the achievable values of all buyers are 6710.78K$, 6667.61K$, 6698.38K$ and
6719.64K$ respectively. All values obtained are more optimized than the GAOPF based
wheeling pricing approach.

If we look at this approach from the perspective of the transmission owner, then the prizes
obtained protect the interests of the transmission owner. If the same approach is taken from the
customer's point of view, the prizes obtained seem to be better than the results obtained from
GAOPF based approach.

Table 4: Wheeling Prices Allocation in IEEE-30 Bus System based on Simultaneous bilateral and multilateral Power Market with Base Case using
Rudnick Tracing and MBA-OPF

Wheeling Pricing Methods ( wheeling prices in k$)

used used
abs ZCF used rev
used used used used
TB1+T TB1+T TB1+T used rev used rev
abs abs ZCF ZCF
S. B2+TB used B2+TB used B2+TB3 TM1 TM2
Use used TM1 TM2 TM1 TM2
N 3 ggdf ZCF 3 ggdf reverse ggdf multilat multilat
rs abs ggdf multilat multilat multilat multilat
o. simulta ggdf simulta ggdf simulta eral eral
eral eral eral eral
neous neous neous ggdf ggdf
ggdf ggdf ggdf ggdf
bilatera bilatera bilateral
l l

1 S1 1532.18 1472.51 1443.54 1481.36 1481.42 1374.00 1353.29 1323.26 1440.466 1417.62 1450.02 1405.59
2 S2 305.40 298.96 301.45 309.35 267.86 257.41 253.91 259.93 232.20 236.94 243.84 241.41
3 S5 97.79 126.64 135.95 139.51 55.59 60.46 55.81 58.47 -2.68 -4.10 -2.36 -2.43
4 S8 168.27 156.81 120.12 123.26 37.85 65.15 64.13 64.84 0.38 -1.23 -0.11 0.53
5 S11 123.44 150.91 146.54 150.38 105.61 87.80 87.88 83.97 29.65 34.70 24.11 30.17
6 S13 176.61 180.88 184.68 189.52 135.66 104.30 104.46 92.57 44.70 57.25 50.37 52.50
Total 2403.69 2386.71 2332.28 2393.38 2084.00 1949.12 1919.47 1883.04 1744.71 1741.19 1765.86 1727.77
7 B2 177.68 177.94 177.35 162.81 93.36 88.85 89.10 87.50 5.36 14.89 11.27 4.30
8 B3 34.88 35.06 34.81 34.22 13.91 13.63 13.59 13.58 -7.30 -6.13 -6.46 -7.52
9 B4 103.45 103.51 103.26 103.33 61.19 60.66 60.37 60.47 20.28 22.55 21.89 20.13
10 B5 1539.22 1513.81 1536.38 1516.20 1325.10 1317.11 1311.20 1311.25 1189.35 1197.27 1186.39 1199.29
11 B7 419.30 414.89 418.53 419.12 320.96 318.93 317.21 317.49 238.69 242.65 239.18 240.48
12 B8 616.34 613.52 615.20 636.21 487.64 481.29 477.94 479.54 368.59 375.12 367.90 371.63
13 B10 173.08 172.03 172.76 174.91 128.22 133.24 132.18 133.03 100.64 93.42 99.78 101.59
14 B12 316.58 316.04 315.99 319.17 194.90 203.32 201.68 203.35 99.31 90.89 99.99 99.97
15 B14 231.38 231.20 230.95 231.90 141.37 147.53 146.30 147.50 70.33 64.17 70.33 70.88
16 B15 289.51 289.22 288.98 290.58 201.99 211.54 209.76 211.49 143.50 131.87 142.27 144.87
17 B16 106.77 106.30 106.57 108.22 76.78 79.79 79.14 79.75 56.84 53.06 56.42 57.37
18 B17 290.78 289.29 290.24 293.44 235.05 243.88 241.89 243.57 210.03 197.48 207.38 212.18
19 B18 121.05 120.74 120.83 121.96 92.96 96.78 95.97 96.73 77.28 72.26 76.28 78.08
20 B19 363.55 362.19 362.88 365.79 308.15 320.05 317.35 319.73 293.25 276.69 288.70 296.42
21 B20 79.85 79.51 79.70 80.42 62.21 64.65 64.11 64.58 52.83 49.51 52.14 53.37
22 B21 653.52 650.87 652.31 657.12 521.09 543.46 538.95 542.77 461.92 429.19 455.46 466.77
23 B23 126.82 126.54 126.59 127.64 93.77 98.98 98.15 98.92 75.96 68.25 74.99 76.74
24 B24 339.34 338.33 338.71 341.77 273.03 291.69 289.23 291.41 258.90 227.12 254.97 261.67
25 B26 163.11 162.92 162.81 164.59 144.14 134.92 133.77 134.75 113.08 134.83 111.28 114.31
26 B29 102.56 102.40 102.37 103.63 81.73 77.82 77.17 77.71 56.70 66.30 55.98 57.29
27 B30 462.01 461.29 461.15 466.61 389.42 373.06 369.91 372.53 301.71 342.50 297.26 304.92
Total 6710.78 6667.61 6698.38 6719.64 5246.97 5301.20 5264.96 5287.63 4187.24 4143.89 4163.40 4224.72

Similarly, if look in Used ZCF method, the results obtained for all the sellers in the pool,
simultaneous bilateral and multilateral TM1 and TM2 are 2084.00K$, 1949.12K$, 1919.47K$
and 1883.04K$ respectively. Similarly, the prizes received for all the buyers are 5246.97K$,
5301.20K$, 5264.96K$ and 5287.63K$ respectively. All the results obtained are also more
optimized than the GAOPF based wheeling pricing approach. The results obtained from the
perspective of competitive power market model also show better market structure.

Now, if look in Used reverse method, the results obtained for all the sellers in the pool,
simultaneous bilateral and multilateral TM1 and TM2 are 1744.71K$, 1741.19K$, 1765.86K$
and 1727.77K$ respectively. Similarly, the prizes received for all the buyers are 4187.24K$,
4143.89K$, 4163.40K$ and 4224.72K$ respectively. All obtained values present more optimized
results than the well-known heuristic technique GAOPF based approach in the competitive
power market. For easy execution of results, different bar charts for sellers and buyers are shown
in Figures 6 and 7, in which the suggested algorithm based results are presented.

Fig.6 Allocation of Wheeling Prices for Sellers on IEEE-30 Bus System using Rudnick Tracing
with MBA-OPF
Fig.7 Allocation of Wheeling Prices for Buyers on IEEE-30 Bus System using Rudnick Tracing
with MBA-OPF

To further discuss the consequences and limitations of various tracing methods, both GAOPF
based approach and MBA based approach are replaced with bialek tracing instead of rudnik
tracing. In the used absolute method with bialek tracing, all values in pool, simultaneous bilateral
and multilateral transactions TM1 and TM2 with total values of 2111.12K$, 1935.58K$,
1918.43K$ and 1964.43K$ are obtained respectively. Similarly, for all buyers, this value is
obtained 5279.68K$, 5326.58K$, 5353.94K$ and 5303.89K$ respectively.

In bialek tracing, Calculations of negative flow are not possible due to the proportional sharing
principal. For this reason, the result for pool, simultaneous bilateral and multilateral transactions
in used ZCF method and used reverse method are similar to the results found in the used absolute
method. Here only the price of positive flow is considered, so results of all methods are obtained
equal. For easy execution of the results, graphs are presented in the form of representation 8, 9
for sellers and buyers.
Table 5: Wheeling Prices Allocation in IEEE-30 Bus System based on Simultaneous bilateral and multilateral Power Market with Base Case using
BialekTracing and GAOPF

Wheeling Pricing Methods ( wheeling prices in k$)

used abs used used used ZCF used used used rev used used
TB1+TB2 abs abs TB1+TB2 ZCF ZCF used TB1+TB2 rev rev
used used
S.N Use +TB3 ggdf TM1 TM2 +TB3 ggdf TM1 TM2 rever +TB3 ggdf TM1 TM2
abs ZCF
o. rs simultaneo multilat multilat simultaneo multilat multilat se simultaneo multilat multilat
ggdf ggdf
us eral eral us eral eral ggdf us eral eral
bilateral ggdf ggdf bilateral ggdf ggdf bilateral ggdf ggdf

1478. 1478. 1478.

1479.30 719.09 661.82 1479.30 719.09 661.82 1479.30 719.09 661.82
1 S1 85 85 85
543.5 543.5 543.5
386.80 782.78 821.12 386.80 782.78 821.12 386.80 782.78 821.12
2 S2 6 6 6

3 S5 22.35 0.00 0.00 0.00 22.35 0.00 0.00 0.00 22.35 0.00 0.00 0.00

4 S8 11.08 2.55 136.87 147.11 11.08 2.55 136.87 147.11 11.08 2.55 136.87 147.11

5 S11 46.44 0.16 127.74 171.01 46.44 0.16 127.74 171.01 46.44 0.16 127.74 171.01

6 S13 8.86 66.76 151.95 163.38 8.86 66.76 151.95 163.38 8.86 66.76 151.95 163.38
2111. 2111. 2111.
Total 1935.58 1918.43 1964.43 1935.58 1918.43 1964.43 1935.58 1918.43 1964.43
12 12 12

7 B2 84.27 66.07 58.13 54.85 84.27 66.07 58.13 54.85 84.27 66.07 58.13 54.85

8 B3 2.42 1.70 2.38 2.33 2.42 1.70 2.38 2.33 2.42 1.70 2.38 2.33

9 B4 42.05 30.33 40.63 39.59 42.05 30.33 40.63 39.59 42.05 30.33 40.63 39.59
2274. 2274. 2274.
2851.84 3125.02 3178.84 2851.84 3125.02 3178.84 2851.84 3125.02 3178.84
10 B5 53 53 53
826.6 826.6 826.6
499.50 770.26 761.45 499.50 770.26 761.45 499.50 770.26 761.45
11 B7 7 7 7
526.6 526.6 526.6
553.65 0.00 0.00 553.65 0.00 0.00 553.65 0.00 0.00
12 B8 7 7 7

13 B10 27.75 44.71 36.60 31.98 27.75 44.71 36.60 31.98 27.75 44.71 36.60 31.98

14 B12 87.62 26.97 33.17 32.02 87.62 26.97 33.17 32.02 87.62 26.97 33.17 32.02

15 B14 37.54 20.06 23.29 22.27 37.54 20.06 23.29 22.27 37.54 20.06 23.29 22.27

16 B15 64.81 29.02 31.15 30.07 64.81 29.02 31.15 30.07 64.81 29.02 31.15 30.07

17 B16 17.29 4.24 6.04 6.03 17.29 4.24 6.04 6.03 17.29 4.24 6.04 6.03
146.2 146.2 146.2
48.14 136.24 131.22 48.14 136.24 131.22 48.14 136.24 131.22
18 B17 3 3 3

19 B18 27.57 6.30 14.56 14.47 27.57 6.30 14.56 14.47 27.57 6.30 14.56 14.47
175.4 175.4 175.4
132.97 187.25 177.47 132.97 187.25 177.47 132.97 187.25 177.47
20 B19 6 6 6

21 B20 5.49 8.75 7.47 6.66 5.49 8.75 7.47 6.66 5.49 8.75 7.47 6.66
310.6 310.6 310.6
582.37 434.78 386.02 582.37 434.78 386.02 582.37 434.78 386.02
22 B21 8 8 8

23 B23 18.83 10.59 9.51 9.33 18.83 10.59 9.51 9.33 18.83 10.59 9.51 9.33

24 B24 95.16 119.14 99.68 92.35 95.16 119.14 99.68 92.35 95.16 119.14 99.68 92.35

25 B26 70.68 44.51 52.05 49.57 70.68 44.51 52.05 49.57 70.68 44.51 52.05 49.57

26 B29 18.75 10.10 12.31 11.99 18.75 10.10 12.31 11.99 18.75 10.10 12.31 11.99
419.2 419.2 419.2
235.63 273.41 265.36 235.63 273.41 265.36 235.63 273.41 265.36
27 B30 1 1 1
5279. 5279. 5279.
Total 5326.58 5353.94 5303.89 5326.58 5353.94 5303.89 5326.58 5353.94 5303.89
68 68 68
 Fig.8 Allocation of Wheeling Prices for Sellers on IEEE-30 Bus System using Bialek Tracing
with GAOPF

Fig.9 Allocation of Wheeling Prices for Buyers on IEEE-30 Bus System using Bialek Tracing
with GAOPF

Just as the GAOPF based approach in rudnick tracing has been compared to the MBA based
approach, the bialek tracing also compares the results from the GAOPF based approach with the
MBA based approach with used absolute method pool, simultaneous bilateral and multilateral
transactions TM1 and TM2, The total values of all the sellers for are 2069.52K$, 1917.82K$,
1895.58K$ and 1954.35K$ respectively.
Table 6: Wheeling Prices Allocation in IEEE-30 Bus System based on Simultaneous bilateral and multilateral Power Market with Base Case using Bialek
Tracing and MBA-OPF

Wheeling Pricing Methods ( wheeling prices in k$)

used abs used ZCF used used used rev

used abs used abs used rev used rev
TB1+TB2 TB1+TB2 ZCF ZCF used TB1+TB2
S. used TM1 TM2 used TM1 TM2
Use +TB3 ggdf +TB3 ggdf TM1 TM2 rever +TB3 ggdf
N abs multilat multilat ZCF multilat multilat
rs simultaneo simultaneo multilat multilat se simultaneo
o. ggdf eral eral ggdf eral eral
us us eral eral ggdf us
ggdf ggdf ggdf ggdf
bilateral bilateral ggdf ggdf bilateral

1449. 1449. 1449.

1465.73 710.52 658.42 1465.73 710.52 658.42 1465.73 710.52 658.42
1 S1 70 70 70
532.8 532.8 532.8
383.25 773.46 816.91 383.25 773.46 816.91 383.25 773.46 816.91
2 S2 4 4 4

3 S5 21.91 0.00 0.00 0.00 21.91 0.00 0.00 0.00 21.91 0.00 0.00 0.00

4 S8 10.86 2.53 135.24 146.35 10.86 2.53 135.24 146.35 10.86 2.53 135.24 146.35

5 S11 45.52 0.16 126.21 170.13 45.52 0.16 126.21 170.13 45.52 0.16 126.21 170.13

6 S13 8.69 66.15 150.14 162.54 8.69 66.15 150.14 162.54 8.69 66.15 150.14 162.54
2069. 2069. 2069.
Total 1917.82 1895.58 1954.35 1917.82 1895.58 1954.35 1917.82 1895.58 1954.35
52 52 52

7 B2 83.44 65.85 57.87 54.81 83.44 65.85 57.87 54.81 83.44 65.85 57.87 54.81

8 B3 2.40 1.70 2.37 2.33 2.40 1.70 2.37 2.33 2.40 1.70 2.37 2.33

9 B4 41.63 30.23 40.45 39.56 41.63 30.23 40.45 39.56 41.63 30.23 40.45 39.56
2252. 2252. 2252.
2842.34 3111.12 3176.39 2842.34 3111.12 3176.39 2842.34 3111.12 3176.39
10 B5 03 03 03
818.5 818.5 818.5
497.83 766.83 760.87 497.83 766.83 760.87 497.83 766.83 760.87
11 B7 0 0 0
521.4 521.4 521.4
551.81 0.00 0.00 551.81 0.00 0.00 551.81 0.00 0.00
12 B8 6 6 6

13 B10 27.48 44.56 36.43 31.96 27.48 44.56 36.43 31.96 27.48 44.56 36.43 31.96

14 B12 86.75 26.88 33.02 31.99 86.75 26.88 33.02 31.99 86.75 26.88 33.02 31.99

15 B14 37.17 19.99 23.19 22.26 37.17 19.99 23.19 22.26 37.17 19.99 23.19 22.26

16 B15 64.17 28.92 31.01 30.05 64.17 28.92 31.01 30.05 64.17 28.92 31.01 30.05

17 B16 17.12 4.23 6.01 6.02 17.12 4.23 6.01 6.02 17.12 4.23 6.01 6.02
144.7 144.7 144.7
47.98 135.63 131.12 47.98 135.63 131.12 47.98 135.63 131.12
18 B17 8 8 8

19 B18 27.30 6.27 14.50 14.46 27.30 6.27 14.50 14.46 27.30 6.27 14.50 14.46
173.7 173.7 173.7
132.53 186.42 177.33 132.53 186.42 177.33 132.53 186.42 177.33
20 B19 2 2 2

21 B20 5.44 8.72 7.44 6.66 5.44 8.72 7.44 6.66 5.44 8.72 7.44 6.66
307.6 307.6 307.6
580.43 432.85 385.72 580.43 432.85 385.72 580.43 432.85 385.72
22 B21 0 0 0

23 B23 18.65 10.55 9.47 9.32 18.65 10.55 9.47 9.32 18.65 10.55 9.47 9.32

24 B24 94.22 118.74 99.24 92.28 94.22 118.74 99.24 92.28 94.22 118.74 99.24 92.28

25 B26 69.98 44.36 51.82 49.54 69.98 44.36 51.82 49.54 69.98 44.36 51.82 49.54

26 B29 18.56 10.07 12.26 11.98 18.56 10.07 12.26 11.98 18.56 10.07 12.26 11.98
415.0 415.0 415.0
234.84 272.20 265.16 234.84 272.20 265.16 234.84 272.20 265.16
27 B30 6 6 6
5227. 5227. 5227.
Total 5308.82 5330.11 5299.80 5308.82 5330.11 5299.80 5308.82 5330.11 5299.80
45 45 45
 Fig.10 Allocation of Wheeling Prices for Sellers on IEEE-30 Bus System using Bialek Tracing
with MBA-OPF

Fig.11 Allocation of Wheeling Prices for Sellers on IEEE-30 Bus System using Bialek Tracing
with MBA-OPF

Similarly, the total values for all the buyers are obtained 5227.45K$, 5308.82K$, 5330.82K$ and
5299.80K$ respectively. All values obtained give more optimized results than values obtained by
GAOPF based approach. Here too, due to calculating only positive flow, the same result is used
in all transactions, like the used absolute method in the used ZCF and used reverse method. The
bar representation for execution of the results is presented in Figures 10 and 11 for all sellers and
buyers.

In order to test the veracity of the results, both tracings (Rudnick Tracing & Bialek tracing) with
both the above mentioned approaches (GAOPF based approach and MBA based approach) and
the three variants (used absolute, used ZCF and used reverse) have been fitted with the Indian
utility 62-bus system in the paper.

Table 7: Comparisons of both optimization techniques with all three pricing methods in Indian
utility-62 bus system using Rudnick Tracing
TC=41214 Rs. in lac Seller share,30% TC=12364.2 Rs. in lac Buyer share ,70%TC=28849.8 Rs.in lac
Wheeling Pricing methods ( wheeling prices Rs. in lac)
Rudnick tracing with GAOPF Rudnick tracing with MBA-OPF
Tracing Seller Buyer Total Seller Buyer Total
used abs 10880.5 32748.64 43629.14 10835.16 32695.07 43530.23

Used abs simultaneous bilateral 11259.66 32476.63 43736.29 11177.24 32315.9 43493.14

Used abs TM1 11333.85 32464.27 43798.12 11255.54 32303.53 43559.07

Used abs TM2 11300.88 32604.4 43905.28 11243.18 32488.99 43732.17

used ZCF 8955.8 26933.35 35889.15 8893.98 26875.65 35769.63

Used ZCF simultaneous bilateral 9400.91 26962.2 36363.11 9264.91 26883.89 36148.8

Used ZCF TM1 9499.83 26945.71 36445.54 9310.24 26859.16 36169.4

Used ZCF TM2 9343.21 26995.17 36338.38 9223.69 26900.38 36124.07

used reverse 7068.2 21204.6 28272.8 7010.5 21056.23 28066.73

Used reverse simultaneous

7224.81 21443.64 28668.45 7117.66 21344.73 28462.39
bilateral

Used reverse TM1 7261.91 21517.83 28779.74 7138.26 21443.64 28581.9

Used reverse TM2 7158.87 21418.92 28577.79 7092.93 21274.67 28367.6

In the said bus system with GAOPF and rudnick tracing, the used absolute method for pool,
simultaneous bilateral and multilateral transactions TM1 and TM2 gives the total values of
10880.5, 11259.66, 11333.85 and 11300.88 Rs. in lac for all sellers respectively. Similarly, for
all buyers, these values are obtained as 32748.64, 32476.63, 32464.27 and 32604.4 Rs. in lac
respectively. In the used ZCF method, the total values for all sellers for pool, simultaneous
bilateral and multilateral transactions TM1 and TM2 yield 8955.8, 9400.91, 9499.83 and 9343.21
Rs. in lac respectively. Similarly, for all buyers, these values are obtained as 26933.35, 26962.2,
26945.71 and 26995.17 Rs. in lac respectively. In addition to the used reverse method, pool,
simultaneous bilateral and multilateral transactions TM1 and TM2 yield total values 7068.2,
7224.81, 7261.91 and 7158.87 Rs. in lac and for all sellers respectively. Similarly, for all buyers,
these values are obtained as 21204.6, 21443.64, 21517.83 and 21418.92 Rs. in lac respectively.

Like the IEEE 30-bus system in the Indian utility 42-bus system, the total result of the sellers and
buyers obtained by the used absolute method is more than the fixed transmission cost. Apart
from this, only the positive flow value in the used ZCF method and in the reverse method used,
customers get incentive. Here if we compare the results obtained from GAOPF with the results
obtained from MBA -OPF, it is known that the results obtained from MBA -OPF are more
optimized and show better results for all transmission entities (as we can see in Table 7).

On the other hand, if we look at the results of bialek tracing and GAOPF, we find that in the used
absolute method for pool, simultaneous bilateral and multilateral transactions TM1 and TM2 the
total values for all sellers are obtained as 8922.83, 9417.4, 9545.16 and 9277.27 Rs. in lac
respectively. Similarly, for all buyers, these values are obtained as 26908.62, 26945.71, 26908.62
and 27024.02 Rs. in lac respectively. Used ZCF and used reverse methods also get the same
results as the used absolute method, just like the IEEE 30-bus system.

Similarly if in bialek tracing also compares the results obtained from GAOPF with the results
obtained from MBA -OPF, it is known that the results obtained from MBA-OPF are more
optimized and show better results for all transmission entities (as we can see in Table 8).

After analyzing all the above results, it comes out that rudnick tracing is more practical for price
calculation for all users. The reason for this is that with this tracing it is possible to interpret both
negative and positive flows, as well as MBA based tracing gives better results to protect the
interests of both the transmission owner and the users.
Table 8: Comparisons of both optimization techniques with all three pricing methods in Indian
utility-62 bus system using Bialek Tracing

TC=41214 Rs. in lac Seller share,30% TC=12364.2 Rs. in lac Buyer share ,70%TC=28849.8 Rs.in lac
Wheeling Pricing methods ( wheeling prices Rs. in lac)
Bialek tracing with GAOPF Bialek tracing with MBA-OPF

Tracing Seller Buyer Total Seller Buyer Total

used abs 8922.83 26908.62 35831.45 8811.55 26817.95 35629.5

Used abs simultaneous bilateral 9417.4 26945.71 36363.11 9264.91 26826.19 36091.1

Used abs TM1 9545.16 26908.62 36453.78 9433.88 26797.34 36231.22

Used abs TM2 9277.27 27024.02 36301.29 9145.39 26916.86 36062.25

used ZCF 8922.83 26908.62 35831.45 8811.55 26817.95 35629.5

Used ZCF simultaneous bilateral 9417.4 26945.71 36363.11 9264.91 26826.19 36091.1

Used ZCF TM1 9545.16 26908.62 36453.78 9433.88 26797.34 36231.22

Used ZCF TM2 9277.27 27024.02 36301.29 9145.39 26916.86 36062.25

used reverse 8922.83 26908.62 35831.45 8811.55 26817.95 35629.5

Used reverse simultaneous bilateral 9417.4 26945.71 36363.11 9264.91 26826.19 36091.1

Used reverse TM1 9545.16 26908.62 36453.78 9433.88 26797.34 36231.22

Used reverse TM2 9277.27 27024.02 36301.29 9145.39 26916.86 36062.25

5 CONCLUSION

Hybrid power market is spreading rapidly. In this power market, knowledge of simultaneous
bilateral and multilateral transactions with pool power market plays an important role in
understanding the power market. If the contribution of each user in the entire power market is
known, then a healthy and clear environment is created in the power market. Actual power flow
based methods in power market are generally reality based methods. Under these different types
of optimization techniques and tracings, trying to basically benefit all the components of the
power market can generate a new scenario in the power market.

This paper explains the applicability and limitations of two well known tracing methods. This
interpretation can prove useful in choosing the best tracing technique for price allocation in the
emerging scenario of competitive power market. In addition, a modified bat algorithm based or
suggested approach that replaces new scales compared to established artificial intelligence
techniques for cost optimization. The suggested method makes the used absolute method
dependent on actual power flow even better from the point of view of all entities of transmission.
In future, using this work with other and better optimization methods, better results can be
achieved.

6 ACKNOWLEDGEMENTS

The authors sincerely acknowledge the Director, MITS, Gwalior, India, to help us carry out this
research work.

7 REFERENCES

[1] Loi Lee Lai, Power System Restructuring and Deregulation. John Wiley and Sons, UK. 2001.
[2] Ezra Hausman, Rick Hornby, and Allison Smith, “Bilateral Contracting in Deregulated
Electricity Markets”, A Report to the American Public Power Association, April 18, 2008.
[3] Gomes, Bruno Andre and Pereira Santos, Modelization of Generation Cost and Demand
Uncertainties in Power System Optimization Problems and in Nodal Marginal Price
Calculations, Faculdade De Engenharia Da Universidade do Porto, 2013.
[4] Upadhyay K.G., Singh S.N., Chauhan D.S. and Srivastava G.S., Wheeling Rates and Pricing
Mechanism: An Overview and Key Issue, International Conference on Computer Application in
Electrical Engineering (CERA 01), University of Roorkee, India, 2002
[5] Rudnick H., Palma R. and Fernandez J.E., Marginal Pricing and Supplement Cost Allocation
in Transmission Open Access, IEEE Transaction on Power Systems, 10: 1125-1132, 1995.
[6] Happ H.H., Cost of Wheeling Methodologies, IEEE Transaction on Power Systems, 9(1):
147-156, 1994.
[7] Shirmohammadi D., Filho X.V., Gorenstin B and Pereira M.V.P., Some Fundamental
Technical Concepts about Cost Based Transmission Pricing, IEEE Trans Power System, 11:
1002-1008, 1996.
[8] R.R. Kovacs, A.L.Leverett, “A Load Flow Based Method For Calculating Embedded,
Incremental and Marginal Cost of Transmission Capacity”, IEEE transaction on power systems,
vol. 9, No. 1, February 1994, pp. 272-278.
[9] Garg N.K., Palwalia D.K. and Sharma Harish, Transmission Pricing Practices: A Review,
IEEE International Conference on Engineering (NUICONE), Ahmadabad, 2013, pp. 1-6.
[10] N. Kumar, Y.R.V. Reddy, D. Das, N.P. Padhy, “Allocation of Transmission Charge by
using MVA-Mile Approaches for Restructured Indian Power Utility”, IEEE Power and Energy
Society General Meeting, 24-29 July 2011, pp.1-6.

[11] Babasaheb Kharbas, Manoj Fozdar, Harpal Tiwari, “Comparative Assessment of MW-mile
and MVA-mile Methods of Transmission Tariff Allocation and Revenue Reconciliation” IEEE
Power & Energy Society General Meeting, 21-25 July 2013, pp.1-5.

[12] Su Ching-Tzong and Liaw Ji-Horng 2001, Power Wheeling Pricing Using Power Tracing
and MVA-KM Method, Proc. Power Tech., IEEE Porto,1: SSM1–212.
[13] L. G. Manescu, D. Rusinaru, P. Dadulescu, and V. Anghelina, Usage Based Allocation for
Transmission Costs under Open Access. In: Proc. of IEEE Bucharest PowerTech Conf.,
Bucharest, Romania, June 28–July 2, 2009.
[14] J. Pan, Y. Teklu, S. Rahman, and K. Jun, Review of Usage-based Transmission Cost
Allocation Methods under Open Access, IEEE Trans Power Syst 15 (2000), 1218–1224.
[15] Ng Wai Y. 1981, Generalized Generation Distribution Factors for Power System Security
Evaluations, IEEE Transaction on Power Apparatus and Systems, PAS 100(3): 1001-1005.
[16] Bialek J., Tracing the Flow of Electricity, IEE Proc. Generation Transmission and
distribution, 143: , 1996, pp. 313-320.
[17] Bialek J., Topological Generation and Load Distribution Factors for Supplement Charge
Allocation in Transmission Open Access, IEEE Trans. Power Syst., 12(3):, 1997, pp.1185-1193.
[18] Oana Pop, Stefan Kilyeni, Petru Andea, Constantin Barbulescu and Cristian Craciun, Power
Flow Tracing Method for Electricity Transmission and Wheeling Pricing, Journal of Sustainable
Energy, 1(4):, 2010, pp. 63-70.
[19] Strbac G., Kirschen D. and Almed S., Allocating Transmission System Usage on the Basis
of Traceable Contributions of Generators and Loads to Flow, IEEE Transactions on Power
Systems, 13(2):, 1998, pp. 527-534.
[20] Kischen D., Allan R. and Strbac G., Contributions of Individual Generators to Loads and
Flow, IEEE Transactions on Power Systems, 12(1):, 1997,pp. 52-66.
[21] Barcia P. and Pestana R. 2010, Tracing the Flows of Electricity, Electrical Power and
Energy Systems, 32(4): 329-332.
[22]https://www.alroomi.org/multimedia/Power_Flow/30BusSystem/
IEEE30BusSystemDATA1.pdf.
[23]https://www.alroomi.org/multimedia/Power_Flow/62BusSystem/
Indian62BusSystem_DATA.rar
[24]shodhganga.inflibnet.ac.in/bitstream/10603/4566/13/13_chapter%203.pdf
[25] Kalyanmoy Deb, “Multi-objective Optimization using Evolutionary algorithms”, john
Wiley and Sons, 2001.
[26] S.N. Sivanandam, S.N. Deepa, “Principles of Soft Computing: Second Edition”,
www.wileyindia.com
[27] Palley, Grunwald & Neuweiler 2001
[28] Xian-Bing Meng , X.Z. Gao , Yu Liu , Hengzhen Zhang, A novel bat algorithm with habitat
selection and Doppler effect in echoes for optimization, Expert Systems with Applications, Vol.
42, 2015, pp. 6350–6364.
[29] R. Gnanadass, N. P. Padhy, “A New Approach for Transmission Embedded Cost Allocation
in Restructured Power Market”, Journal of Energy & Environment 4 (2005) 37 – 47.

8 APPENDIX A: Modified BA and basic BA OPF-

Modified Bat Algorithm Inputs- Main parameters of standard BA

α-- 0.99, ϒ-- 0.9, fmin –0, fmax—1.5, Ao—[1, 2], Yo—[0, 1], M—Maximum no. of iteration=10, N
—No. of individuals i.e. bats (population size)

Additional input parameters proposed in MBA are:-

G--Frequency of updating the loudness and pulse emission rate G=10.

w-- Inertia weight to adjust velocity [0.5, 0.9].

C- - Compensation rate to consider Doppler Effect in echoes [0.1, 0.9]

P- - Probability of habitat selection [0.6, 0.9]

Q- - Contraction-expansion coefficient, to update position of bats quantum behaviors [0.5, 1].

STEP 1:- For iteration t=0, initialize the population and related parameters.

STEP 2:- Evaluate the objective function and determine fitness of each individual.
STEP 3:- For iteration in the maximum count (i.e. for t < m), invoke habitat selection, by

comparing probability of habitat selection (i.e. P) with a uniform random no. rand1 [0, 1].

Step 3.1—If random no. in [0,1] is smaller than P, bats will choose the quantum behavior to
forage in a wide range of habitats, that means generate new solutions using equation (9) as
follows:-

, if rand (0, 1) < 0.5,

(9)

, otherwise

Where, is the global best solution.

Updated off springs after tth iteration

is a random no. uniformly distributed between 0 & 1 to avoid the case of zero-
division error
is the contraction-expansion coefficient.

Step 3.2—If random no. in [0,1] is not smaller than P, bat will choose mechanical behavior to
forage in habitat, however due to inclusion of Doppler effect the new off springs and at
time step t will be determined by equation (10), (11), (12) and (13), as follows:-

(10)

(11)

(12)

(13)

Where, C—Wave speed in medium= 340m/s speed of sound in air

 -- Velocity of individual at line step‘t’ in a D-dimension and iϵ [1,…..N], jϵ [1,…….D]

-- Velocity of best found individual in time step t and in D dimensional search space or
velocity corresponding to global best position

-- Compensation rate of ith individual. It is taken as a +ve volume between 0& 1.

w—Inertia weight to update velocity. It is a uniform random no.

ϵ -- A smallest constant in computer to avoid the possibility of division by zero

is the updated off spring means new solution

STEP 4:- Implement the local search by considering the loudness of surrounding environment
apart from loudness produced by the bat itself.

NOTE: - It is known that as soon as bat approaches the prey, its loudness is bed and pulse
emission rate is increased. Thus bat approaches the prey as silent as possible. But apart from the
bat’s loudness (i.e loudness induced by the bats themselves), the loudness produced by the other
things in the surrounding environment is also present. So to make the local search more efficient
the relative loudness between the loudness of a specific bat and the arrange loudness of all the
bats is regarded as are influence factor in the local search of the bat for the prey nearby

Step 4.1:- Compare a random no. pulse emission rate. If random no. bi (0, 1) is larger than ϒi

then generate a local solution around the selected best solutions equation (14) and (15)

(14)

(15)

Where is a Gaussian distribution mean 0 and standard deviation .

is the arrange loudness of all the bats at time step t.

is a small +ve no. used to ensure .

 is a global best solution.

STEP 5:- evaluate the fitness of each individual.

STEP 6:- Update solutions, loudness and pulse emission rate globally, in the same manner as
that of standard BA using equation (16), (17), (18), (19).

If (16)

(17)

(18)

(19)

STEP 7:- Rank the solution and find the current best .

STEP 8:- If does not improve in G time steps, means when all the bats do not find a better
prey than the previous one in several time steps G, this is the situation of being trapped in local
optima. Then the bats would fly to another place to search for prey. Mathematically it can be
formulated as, re-initialize the loudness Ai and set pulse emission rate ri, temporarily to a high
value, so that bats can search globally.

STEP 9:- increase the iteration count by 1.

STEP 10:- if termination criteria is met, finish the algorithm otherwise go to step 3.

STEP 11:- Obtain the individual the best objective function value in the population.

Basic BAT Algorithm:-

STEP 1:- Initialize bat population (i.e. possible solutions) for each bat (individual) initialize the
position (i=1,….N) and velocity can be set to zero or any other small value. Also determine
pulse frequency and using equation (20).
 (20)

Where, is a random vector taken from a uniform distribution. The valuer of and
depends upon problem domain size. Here and .

Now initialize pulse emission rate ri and loudness of sounds Ai.

STEP 2:-For (t < m) the iteration in the maximum count i.e. M, generate new solutions by

adjusting frequency using (1) and updating locations/positions and velocities using (21) & (22).

(21)

(22)

Here is the current global solution at t th iteration, which is located after comparing all the
solutions among all the bats.

STEP 3:-Apply local search by comparing pulse emission rate ri with a random number rand 1.

If the pulse rate ri is lesser enough than the random no. rand 1 local search is invoked. Smaller
pulse rate indicates that the optimum solution is nearby. Hence local search can provide the best
solution. Local search strategy is best on the equation (24).

If
(23)

(24)

Where is the average loudness at tth iteration or at time t, of all the individuals/bats and
is a random vector again drawn/taken from a uniform distribution.

Thus generate a local solution by evaluate a fitness around the selected best solution.

STEP 4:- Update solutions by flying randomly to exploit a global search is achieved by
updating/ controlling loudness and pulse rates using equation (27) & (28), as follows:-
If (25)

(26)

(27)

(28)

Or (29)

Where and are constants. In this implementation α=0.99, ϒ=0.9 has been
taken, while initial loudness and pulse emission rate are taken as uniformly distributed in
the interval [1, 2] and [0, 1] respectively.

STEP 5:- Rank each individual based on fitness value and save the best i.e. update the current
global best solution.

STEP 6:- If termination criteria is met then finish the algorithm otherwise go to step 2.

9 APPENDIX B: Rudnick Tracing- For proper allocation of wheeling prices with transacted
power flow, Rudnick awards a very pragmatic approach of power tracing. This tracing method
depends on three generalized distribution factors as Generalized Shift Distribution Factors
(GSDF), Generalized Generation Distribution Factors (GGDF) and Generalized Load
Distribution Factors (GLDF). GSDF represents the incremental change in any line by changing
the power injection in any bus, GGDF symbolizes the effect of change in generation of any bus
on the power flows over any line and GLDF characterizes the effect of change in any load bus on
to the power flow over any line [14, 15].

(30)

where and are the reactance matrix elements and is line reactance. The generalized
generation distribution factor represents the generator supply allocation by generator i in line i-j
due to gth generator as:
(31)

(32)

Where, (AF)ij is actual apparent power flow over line i-j, Sg is generation on gth bus, R is
reference bus and N is total number of generator buses.

GLDF determines the amount of load supplied by lth load bus, which flows over a particular line
i-j and can be displayed as follows:

(33)

(34)

where, Sl is load on lth load bus and M is the total number of load buses.

APPENDIX C: Bialek Tracing- Bialek tracing is based on Proportional Sharing Principle

(PSP). According to PSP, each outgoing flow through every node is in proportion to every
incoming flow at that node. Bialek tracing proposes two algorithms on the basis of Proportional
Sharing Principle [17, 18].

Upstream algorithm is used to determine allocation of power generation at any generator bus.
The total apparent power flow (AF)i through node j may be expressed as :-

(35)
where u(i) is the set of nodes supplying direct node i.(AF)ij is the apparent power flow toward
node i in line i-j and (AF)Gi is the generation of node i and n shows total number of buses. If we
consider all the inflows, nodal flow at node j by is

(36)

Then we get

(37)

Or

(38)

Or

(39)

Where Au is the upstream distribution matrix, (AF) is vector nodal flow and (AF)G is vector of
generations. Here Au is

Where,

(40)

If Au-1 exists then

(41)

Its ith element will be

(42)

By using Proportional Sharing Principle

 (43)

Taking the value of (AF)i from equations (42) to (43) ,we get,

(44)

Or

(45)

where d(i) is the set of nodes supplied directly from node i and Dij,gG is topological distribution
factor which is used to allocate generation due to gth generator. Similarly, for downstream

(46)

where Dij,yL is the topological distribution factor allocated yth load in line i-j.