1

COMPARISON OF ALGORITHMS FOR

LOSS ALLOCATION IN TRANSMISSION
NETWORKS CONSIDERING POWER
REDISPATCH
Delberis A. Lima, and Antonio Padilha-Feltrin, 0HPEHU,(((

due to quite significant alterations of network parameters;

$EVWUDFW$QDQDO\VLVRIWKHSHUIRUPDQFHVRIWKUHHLPSRUWDQW provide effective incentives or disincentives for generators and
PHWKRGVIRUJHQHUDWRUVDQGORDGVORVVDOORFDWLRQLVSUHVHQWHG7KH loads regarding their relative location in the network and their
GLVFXVVHGPHWKRGVDUHEDVHGRQSURUDWDWHFKQLTXHEDVHGRQWKH magnitudes; and finally, it is desirable be simple to understand
LQFUHPHQWDO WHFKQLTXH DQG EDVHG RQ PDWULFHV RI FLUFXLW 7KH and implement.
DOJRULWKPVDUHWHVWHGFRQVLGHULQJGLIIHUHQWJHQHUDWLRQFRQGLWLRQV Several methods have been proposed in the last years,
XVLQJ D NQRZQ HOHFWULF SRZHU V\VWHP ,(((  EXV 3UHVHQWHG
among them we stand out the following: methods based on
DQG GLVFXVVHG UHVXOWV YHULI\ WKH ORFDWLRQ DQG WKH PDJQLWXGH RI
JHQHUDWRUVDQGORDGVWKHSRVVLELOLW\WRKDYHDJHQWVZHOORUSRRUO\ network matrices [7], methods based on proportional sharing
ORFDWHG LQ HDFK QHWZRUN FRQILJXUDWLRQ WKH GLVFULPLQDWRU\ [8], methods based on incremental transmission losses [9],
EHKDYLRU FRQVLGHULQJ YDULDWLRQV LQ WKH SRZHU IORZ LQ WKH methods based on the Pro-Rata technique [10] and methods
WUDQVPLVVLRQOLQHV based on bilateral contracts [11]. A comparative revision of
some methods is made in [12] and an analysis of results is
.H\ :RUGV ORVV DOORFDWLRQ SRZHU IORZ UHGLVSDWFK RI DFWLYH presented in [2] and [13]. This paper presents a performance
SRZHUWUDQVPLVVLRQORVVHV analysis of some of the most important methods for loss
allocation of generators and loads. Results, presented and
I. INTRODUCTION discussed in the paper, consider the position and the magnitude

D ifferent proposals for loss allocation in electric networks

have appeared in last years, for transmission and
distribution systems. Obviously, the larger focus has been the
of generators, and the possibility of having agents well or
poorly positioned in the network. In addition, tests aimed at
verifying the methods sensibility considering redispatch of
transmission due to the reached values and the great amount of active power, are performed. Thus, this work intends to
agents with open access to the transmission networks. In the provide a contribution to indicate problems and solutions for a
Brazilian South-Southeastern network, the cost of the losses good evaluation and application of loss allocation methods.
can reach values up to one billion of Reais [1], consequently it
II. LOSS ALLOCATION METHODS
is necessary to develop more and better methods. In this way,
several proposals have been appearing and some comparative Some studies have been indicating that the methods based
studies were presented [2], [3]. For distribution systems some on network matrices as well as the methods based on
proposals are made, mainly, considering the presence of incremental transmission losses coefficients are the categories
distributed generation [4]-[6]. with larger potentialities and most suitable for solving the loss
Loss allocation methods are aimed at assigning to each allocation problem [9], [14]. The proposal in [14] is, in part,
generator and load, the responsibility of paying for part of the the result of a taskforce coordinated by the MAE (market
system transmission total losses. Therefore, some wholesale of energy - Brazil). For that reason, this paper has as
characteristics are important for a suitable method [7], for objective to compare those two categories with the pro-rata
example: reflect the magnitude of the power injections at each technique that now has still been used in some countries,
bus; consider the relative position of a bus in the transmission including Brazil. To represent the described categories the
network; be consistent with the solved power flow; be stable following methods were chosen:

This work was, in part, financed by CTEEP - Company of Transmission ΠZ-bus - method based on the Z-bus matrix [7];
of electric power of São Paulo. Œ AIL – Based on the incremental loss factors [14];
D. Lima and A. Padilha-Feltrin are with the Department of Electrical
Engineering, Universidade Estadual Paulista, Campus de Ilha Solteira. Ilha
ΠPro-Rata (P) - method that use the Pro-Rata technique
Soltiera – SP – Brazil (e-mail: [email protected], based on active power [10].
[email protected]).
.
 2

$=EXV0HWKRG The second term of the expression represents the sum of the
This method presents a different and interesting power flow connected through the matrix (Z-bus) to the bus N.
characteristic that is the fact of exploring the equations of It can be verified that this sum is equal to zero. Consequently,
electric circuits without any simplification. The idea of the the first term of the right-side expression represents the system
method is to divide the system losses (/) among the "Q" buses losses (/):
ÑÔ !#" * Ë !#" Ûá
/ = §ÒÊ , % ÌÌ Ê 5% $ , $ ÜÜ â
Ô
of the network, from the power flow solution:

/ = Ê /

Ô ÝÔ
(7)
(1)
Ó 1 Í 1 ã
% $


(/ )
1

The loss component is the fraction of the system

The system losses shown through the resistance matrix “5´
losses allocated for the real power injection at the bus "N". can be decomposed in “QE” portions of losses that represent
Equation (1) is solved from the admittance matrix (< *M%) the losses attributed to each bus. Then:
ÑÔ Ë '( ÛáÔ
and the vector of complex bus voltage ((), or using the
/* = §Ò , * * ÌÌ Ê *) ) Ü
impedance matrix (= 5M;) and the vector of complex bus
& 5 , Üâ
Ô ÝãÔ
(8)
Ó Í 1
current injection (,). Considering the losses equation: )

/ = § Ò Ê ( ,  * â
Ñ á

It can be seen in [7], that the losses components (/ ),
Ó 1 ã
(2)
  represent the connections among the currents injections in the
“QE´ buses of the system with the current injection of the bus
N.
Where:
( – vector of complex bus voltages.
, - vector of complex bus current injections. %$,/0HWKRG
§- Real part of the expression. /RVV0RGHO

Since the current injection at bus N is obtained by: The AIL method is based on Incremental Transmission

,
= Ê <
(
 Losses (ITL) factors, that are defined through the variation of
(3) the electric losses when the power injected in each bus is

1 marginally increased [15]. Here, the lineal model will be used
The losses may be expressed by the admittance matrix for obtaining the losses factor as was done in [14]. The total
< *M%. losses can be calculated by:
ÑÔ  Ë Ûá
/ = §ÒÊ ( Ì Ê <  * ( * ÜÔ /539:5 = ) 5)


ÜâÔ
687 5

Ô Ì  1 (4) ( 9)
Ó 1 Í Ýã


where ) is the vector (Q [) containing the active flows of the

The procedure of loss allocation performed by using the circuits, and 5 is the diagonal matrix (Q [Q ) of resistance of
admittance matrix (<) does not produce good results due to the the lines. The ITL factor can be obtained by calculating the
small number of buses connections present in the admittance sensibility of the expression (9) respect to the bus active
matrix (<). Thus, the electric losses can be represented by the injections:
Z-bus matrix formulation:

ÑÔ  * Ë  ÛáÔ Î ›/13241 Þ Î ›/13241 ›) Þ

/ = § ÒÊ  , ÌÊ
.0/ .0/

 Ì   Ü F=Ï ß +-, = Ï ß = 2 )R5b

 = , Üâ
Ð ›3 à Ð ›) ›3 à +-,
* (10)
Ô ÝÔ
(5)
Ó 1 Í 1 ã
 

Where: Where b is the matrix (Q [Q ) with the sensitivities of the

Z – Z-bus matrix of the network. circuit flows respect to the buses injections. The term f is
With (5) it is possible to obtain an expression with two called LQLWLDOORVV IDFWRUof the bus L, and it can be interpreted
summation terms. One involves the resistance matrix 5 and the as:
other involves the reactance matrix ;.

ÑÔ  * Ë  ÛáÔ ÑÔ  * Ë  Ûá

/ = §Ò Ê Ê Ê ,  ÌÌ Ê
 Ì   , Ü + §
Ô
, Ì 5 Ü â Ò ;   ,  ÜÜâ (6) F; =
Variation in the losses
Ô
Ó 1 Í 1 ÝãÔ Ô
Ó 1 Í 1 ÝÔ
(11)
ã
   
   
Variation in the Injection of the bus i
Where:
5 – Real part of the Z-bus matrix.
; – Imaginary part of the Z-bus matrix.
 3

6KDULQJ/RVVHV In general s@1, consequently the final loss factor is:
=s *F
Y[Z \^]3_
F " (22)
As described in [9] it is necessary to establish a strategy to
divide the losses between generators and loads.
&3UR5DWD0HWKRG3
The system losses around the operation point (3R) can be
approximated using the following expression: The Pro-Rata method for loss allocation can be based on
the power injection or current injection for each bus [10]. To
/?3@:? = /?3@:? ( 3R) + F[ 3 - 3R]
=8> =8>
(12) distribute the system electric losses using the Pro-Rata (P)
technique the following equation should be used:
Under the DC power flow the sum of the all buses
injections is zero. Thus f 3 = 0 and (12) can be written as: 3b
/b =
( 3R) + [F - F A ][ 3 - 3R] Ề
a (23)
= 3b
B8C B8C
/D3E:D /D3E:D (13) b
1

This equation is used in two operating points: 3 3* (bus Where /< is the value of losses allocated for the bus L and
generation vector) and 3 3/ (bus load vector). Considering /3 < / is the injection active power in the busL.
an equal division (50/50) between generators and loads, both
results should be compared, then we obtain: III. RESULTS

.= F[3* + 3/ ] $/RVV$OORFDWLRQ
1 (14)
*+/ An analysis of the performances of loss allocation methods
Where * and / are the system’s total generation and load, considering different situations of power dispatch, comparing
respectively. different methods, is presented. All tested methods allocated
losses equal to zero for buses that do not had generators
)LQDOORVVIDFWRU neither loads. The Pro-Rata (P) method always allocates a part
The loss factors as presented previously are highly of losses for buses with generators and loads, which means that
dependent on the reference bus. Then, it is necessary to define all generators and loads should “ pay” for electrical losses. The
the loss factor for eliminating this dependence AIL and Z-bus allocate positive and negative values of losses,
F* = F - F F
which means that generators and loads can “ pay” (positive
(15) allocation) or “ receive” (negative allocation) for electric
Substituting (15) into (13), and expanding the products of losses.
the vectors, it is obtained: When the negative allocation occurs for a bus, it means that

/J3K:J = - /J3K:J ( 3R) + F * 31 + ... + F* 3G

H8I H8I this bus should receive monetary incentives, and obviously the
(16)
other buses of the system should contribute in order to pay
these incentives.
The part of losses attributed for a bus L is given by f< * 3 < , œ The analyses will be made with the system IEEE 14 [7],
L8M
/N3O:N ( 3R)
r =1-
whose unifilar diagram is shown in Fig. 1. This is a very
(17)
F* 3 useful system, and it allows a good interpretation of
results.
Then, considering 3 3R:
F* 3 = 2 * /R3S:R ( 3R)
P8Q
(18)
Therefore, a loss factor can be defined as follows:

F" = r * F* =
1 *
F (19)
2
To obtain the value of loss allocation for each bus
(generator/load), the injected active power in each bus should
be multiplied by a factor defined as:
$)T = [1 - F"T ] (20)

Due to the DC power flow, it is necessary to apply a

correction factor. The correction factor can be calculated as:
U4V
/W3X4W
s = " (21)
F3
Fig. 1 – IEEE-14 bus system.
 4

Tables I, II and III show the obtained results for the IEEE- objective is to verify the cases where generators and loads are
14 system (base case) and two modified cases, inserting a better distributed in the system. A change in the generation of
generator of 50 MW and 100 MW at bus 8, respectively. Costs bus 1 (slack bus) appeared due to bus 8. This can be verified in
are presented in ($) which is obtained by multiplying the Table II, causing a significant decrease in the system total
allocated losses by 50. This procedure was adopted here in losses, compared to the base case. For the base case the cost of
order to facilitate the comparison with the reference [7]. Many total losses was 677.7 ($) and for the modified case 1 was only
times the losses allocation is presented after attribution of a 448.5 ($). This variation was produced because of the
percentage of the losses for generators and loads, for example, insertion of a generator in bus 8. The method Z-bus and the
50% for loads and 50% for generators, [1]. In this paper this method AIL presented the largest sensibility respect to the
procedure is not adopted, however there are not difficulties to change of generation, attributing a negative value of loss
apply it here with the methods presented. By the results allocation to bus 8 for the benefit it brought to the system (loss
obtained in Table I can be observed that all the losses reduction) with the addition of the generator. It can be noticed
allocation methods attributes the largest value of losses for bus that:
1 (which have the largest output power generation of the ΠLosses allocated with Z-bus and AIL show that bus
system), in the same way bus 3 (which have the largest load) 8 receives incentives.
pay the largest value of loss allocation among the loads. ΠEven with the incentive for bus 8, the method Z-
bus allocated a smaller value (regarding Table I)
TABLE I for all buses of the system. It shows a non-
COSTS OF LOSSES ALLOCATED IN THE SYSTEM IEEE 14 ( BASE CASE)
discriminatory procedure, because the losses were
Bus
reduced and all the buses obtained benefits.
PRO-RATA (P) AIL Z-bus This monetary incentive for bus 8 (Table II), attributed by
nº ($) ($) ($) Z-bus and AIL, is not verified by the Pro-Rata method (P).
1 322.87 335.50 381.97 The Pro-Rata (P) method increased the attribution of losses
2 25.40 1.70 7.98 cost for bus 8. The allocation costs using the Pro-Rata (P)
3 130.73 164.00 139.49 method can be considered discriminatory, since the decrease
4 66.33 57.40 42.47
of losses brought benefits for all bus of the system, except for
5 10.55 6.10 3.56
6 15.54 8.70 23.88 bus 8. Though this is a particular and a didactic case, it shows
7 0.00 0.00 0.00 that some methods can make discriminatory allocations.
8 0.14 0.10 1.16
9 40.94 36.20 25.78 TABLE III
10 12.49 11.70 8.73 COSTS OF LOSSES ALLOCATED IN THE SYSTEM IEEE 14 - MODIFIED CASE 2
(ADDING 100 MW IN THE BUS 8).
11 4.86 4.00 2.54
12 8.47 7.30 4.93
Bus
13 18.73 18.10 12.78
PRO-RATA (P) AIL Z-bus
14 20.68 26.8 22.45
nº ($) ($) ($)
1 80.26 143.70 115.95
2 11.73 6.40 4.12
TABLE II
3 60.36 117.40 123.58
COSTS OF LOSSES ALLOCATED IN THE SYSTEM IEEE 14 - MODIFIED CASE 1
(ADDING 50 MW IN THE BUS 8). 4 30.63 8.90 12.51
5 4.87 0.60 1.00
Bus 6 7.18 3.90 22.83
PRO-RATA (P) AIL Z-bus 7 0.00 0.00 0.00
nº ($) ($) ($) 8 64.01 -3.60 -9.17
1 165.14 240.00 227.67 9 18.90 -0.90 3.03
2 16.97 5.30 5.56 10 5.77 1.50 2.93
3 87.36 131.70 131.16 11 2.24 1.20 1.28
4 44.33 29.60 27.15 12 3.91 4.60 4.64
5 7.05 2.80 2.24 13 8.65 11.00 10.83
6 10.39 5.40 21.97 14 9.55 13.3 14.52
7 0.00 0.00 0.00
8 46.28 -27.60 -24.11
To verify the performance of the methods in a second
9 27.36 15.50 14.13
10 8.35 5.90 5.75
scenario of generation at bus 8, it was considered a power
11 3.25 2.30 1.91 output of 100MW (Table III). For the allocation shown in
12 5.66 5.50 4.81 Table III, the total cost of losses decreased even more and
13 12.52 13.40 11.84 reached the value of 308.0 ($). The Z-bus and the AIL
14 13.82 18.7 18.38 methods continued indicating a negative value for bus 8. Since
the total losses cost decreased, one may expect a larger
Table II presents results inserting a generator of 50 MW at incentive, however the monetary incentive was smaller than
bus 8. Loads and other parameters remained the same. The the previous case. To explain this fact, it should be taken into
 5

account the change in the operation point of the system and lines, considering case A4, are larger than those for case A2.
consequently the variations in the power flows through the In addition, is verified that, in relation to the base case,
transmission lines. Therefore, in this paper, the concept of appears an inversion in the power flows directions through the
counter flow is presented. transmission lines, which, according to the concept of counter
%)ORZLQWKHOLQHV flow [16], indicates a dominant flow leaving bus 8. This fact
indicates a poor positioning of this bus in the system (for this
The results presented in Tables II and III showed a
operation point) and this is reflected in the methods Z-bus and
controversial result from the point of view of electrical losses:
AIL.
even increasing the power generation output in bus 8, which
causes smaller electrical losses, the allocated losses were
worse than considering a generation that causes a larger value 180
of losses. This fact can be explained using the concept of 160

Cost of Losses at the bus 8 ($)

140
counter flow [16]. 120
z-bus
fatloc
The concept of counter flow proposes that an analysis of 100
loss allocation results should verify the intensity variations and 80
the flow direction in the transmission lines. For this analysis, 60
40
are studied cases where generation in bus 8 is increased, as
20
shown in Table IV. Fig. 2 presents the system losses 0
considering variations of power dispatch at bus 8. Fig. 3 -20 0 50 100 150 200 250
presents the cost of losses attributed by the methods Z-bus and -40

AIL to the bus 8 for the variations presented. Power at the bus 8 (MW)

TABLE IV
CASES TO ANALYZE THE FLOW IN THE NETWORK Fig. 3 – Cost of allocated losses to the bus 8 for different power generation
outputs at bus 8.
Cases Variations in the bus 8

A1 Generation of 0 MW
TABLE V
A2 Generation of 50 MW FLOW THROUGH THE TRANSMISSION LINES
Casos A1 A2 A3 A4 A5
A3 Generation of 100 MW
Fluxo na Linha 4-7 (MW) 28.2 -4.2 -36.2 -68.3 -100.3
A4 Generation of 150 MW Fluxo na Linha 4-9 (MW) 15.8 7.3 -1.0 -9.4 -17.8
Fluxo na Linha 6-11 (MW) 7.5 1.9 -3.9 -9.4 -14.6
A5 Generation of 200 MW Fluxo na Linha 13-14 (MW) 6.3 2.5 -1.1 -4.5 -7.7

For the cases presented in Table IV, case A4 reveals the Regarding these results, it can be studied in a more efficient
smallest value of total losses (Fig. 2), however the costs way, the insertion of new generators in the system. Presented
allocated to bus 8, considering case A2, are the smallest ones results indicate that inserting a generator at bus 8 of 50 MW is
(Fig. 3). This happens due to the variations of power flow in better than one of 100 MW, for the system as a whole.
the transmission lines, as shown in Table V.

IV. CONCLUSION
800 This paper presented results with three important methods
700 of loss allocation used in transmission networks, and made a
Cost of Total Losses

600 critical analysis. All the methods did not present problems for
Losses
500 understanding and implementation. The Pro-Rata (P) method
400
does not depend on the network, which represents a drawback.
300
The methods Z-bus and AIL present both positive and negative
200
allocation. The fact of presenting negative allocation can be
100
considered as an advantage. The negative allocation, by the
0
methods AIL and Z-bus, informs which buses (loads and
0 50 100 150 200 250 generators) can be considered as well positioned in the
Power at the bus 8 network, and this can serve as an indicative for increasing the
Fig. 2 – Cost of losses of the system for the different power generation generation or the demand, in certain areas. Obviously this
outputs at bus 8. indication of good places for new generators (or loads) should
be considered evaluating the loss allocation with load curves
Table V shows that the flows presented in the transmission (daily, monthly, seasonally).
 6

A methodology of loss allocation, using conventional Faculty at ECE Department of University of Wisconsin –
techniques associated with counter flow concepts, was Madison. His main interests are in analysis and control of
presented in this paper. This methodology presents as an power systems. (UNESP – Ilha Solteira, Department of
advantage the fact of allowing a fairer allocation and with a Electrical Engineering, Av. Brasil, 56 Caixa Postal 31, 15385-
larger information about transmission networks procedures. 000 Ilha Solteira, SP, Brazil).
e-mail: [email protected]
REFERENCES

[1] Leite da Silva, A M., Costa, J.G.C, Mello, J.C.O, Abreu, J.C.,
Romero,S.P., Treistman,R. Determinação Dos Fatores de Perdas
Aplicados na Medição do mercado Atacadista de Energia Elétrica. XVI
SNPTEE, Outubro de 2001, artigo GAT-018.
[2] A.J. Conejo, J.M. Arroyo, N. Alguacil and A.L. Guijarro, Transmission
Loss Allocation: A comparison of different Practical Algorithms, c d8d8d
egf^hgigj3kmlniporq-sutvfxwyzj{ tv|}j^~ k
vol. 17, no. 1, pp. 571-576, August 2002
[3] D.A. Lima, A. Padilha-Feltrin, Allocation of the costs of transmission
losses. Electricity Power System Research (to appear in 2004).
[4] J. Mutale, G. Strbac, S. Curcic, N. Jenkins, “ Allocation of losses in
distribution systems with embedded generation” , IEE Proc. Generation
Transmission Distribution, vol. 147, no 1, pp. 7-14, Jan. 2000.
[5] P.M. Costa; M.A. Matos “ Loss allocation in distribution networks with
embedded generation” . IEEE Trans. on Power Systems, vol. 19 no. 1,
pp. 384 - 389, Feb. 2004.
[6] I.F.E.D. Denis, “ Methods of losses allocation in distributions systems
with distributed generation” . Phd, Dept. of Electrical Engineering.,
Unesp, Ilha Solteira, 2003.
[7] A. J. Conejo, F. D. Galiana, and I. Kockar, "Z-bus loss allocation," c d8d8d
egf^hgigj3k-orq-sutvfxwyzj{ tv|
, vol. 16, no.1,pp. 105-110, Feb. 2001.
ofq€tvtv‚„ƒ i…mj
[8] J. Bialek,"Tracing the flow of electricity," c d8d
† tvigtvf^hz{ ƒ q-i
,
egf^hgigj^|‡ƒ j^jƒ q-iˆhgig‚Š‰‹ƒ j{ fƒ Œm{ ƒ q-ig~
, vol. 143, no. 4, pp. 313-
320, July 1996.
[9] F.D. Galiana, A.J. Conejo and I. Kockar, “ Incremental Transmission
egf^hgigj^hm€#{ ƒ q-igjˆq-i
Loss Allocation in a Transaction Framework” , c d8d8d
orq-sutvfxwyzj{ tv|}j
, vol 17, no. 1,pp:26-33, Feb. 2002.
[10] J.J Gonzalez and P. Basagoiti, “ Spanish power exchange market and
information system. Design concepts, and operating experience,” in
ofq€tvtv‚„ƒ i…Žq { ‘gt’-“g“g“ orq-sutvf ig‚mj{ f y•”uq-|0–xm{ tvf}—„–g–r˜ ƒ €hz{ ƒ q-igj
c d8d8d c
”uq-i #tvf^tvim€t
, Santa Clara, USA, May 1999, pp. 245-252.
[11] F.D. Galiana and M. Phelan, “ Allocation of transmission losses to
egf^hgigj3k™lni
bilateral contracts in a competitive environment,” c d8d8d
orq-sutvfxwyzj{ tv|}j^~
vol. 15, no. 1,pp. 143-150, Feb. 2000.
[12] T.S.P. Fernandes and K.C.Almeida, “ Methodologies for Loss and Line
Flow Allocation under a Pool-Bilateral Market” , in Proc. 2002 Of 14th
PSCC, Session 23, Paper2, pp-1-7.
[13] D. Lima, "Losses Allocation in Transmission Networks," Master' s
degree dissertation, Dept. of Electrical Engineering., Unesp, ha Solteira,
2003.
[14] A. M. Leite da Silva and J. G. C. Costa, “ Transmission loss allocation –
egf^hgigj3k:orq-sutvfpwyzj{ k ~
Parte I: Single energy markets” , c d8d8d vol. 18, pp.
1389-1394, Nov. 2003.
˜ t#€#{ fƒ € igtvf[…yšwyzj{ tv|}j}eg‘gt-q-f y:›g—ui iz{ fq-‚mg€#{ ƒ q-i„›
[15] O.I. Elgerd, d d c McGraw
Hill, 1982, ch.8.
[16] G. Gross and Shu Tao, “ A Physical-Flow-Based Approach to
Allocating Transmission Losses in a Transaction Framework” , c d8d8d
egf^hgigj3kmlniporq-sutvfxwyzj{ tv|}j^~ k
vol. 15, no. 2, pp. 631-637, May 2000

BIOGRAFIES
'HOEHULV $UDXMR /LPD: He received the B.Sc. degree from
UNESP (2000), the M.Sc (2003) from UNESP. He is currently
doctoring student at UNESP.
e-mail:[email protected]

$QWRQLR 3DGLOKD )HOWULQ 0¶ obtained the B.Sc. degree

from UNIFEI, the M.Sc. and Ph.D. from UNICAMP, Brazil.
He is currently an Associate Professor at UNESP – Ilha
Solteira, SP, Brazil. From 1995 to 1997 he was a Visiting