TRANSPORTATION RESEARCH RECORD 1344 31

Standardization of Pavement Management

Systems in Brazil and Other Developing
Countries
CESAR QUEIROZ, w. RONALD HUDSON, AND RALPH HAAS

The deteriorating condition of paved road networks and the lim- to identify the program that will yield the greatest benefit
ited resources available for rehabilitating these roads in devel- over the selected analysis period. At the project level, detailed
oping countries underscore the need for more rational approaches consideration is given to alternative design, construction,
to select priority links on a road network. As a result, some
maintenance, and rehabilitation activities for a particular sec-
developing countries have established pavement management
systems (PMSs) to better manage their road infrastructure assets. tion or project within the network so that an optimum strategy
A main objective of a PMS is to use reliable information and can be identified (2).
decision criteria in an organized framework to produce a cost- Pavement management was developed in the United States
effective pavement program. Pavement management was devel- and Canada and has been widely applied in North America,
oped in the United States and Canada and has been widely applied but there is a tremendous benefit to be gained by applying
in North America, but there is a tremendous benefit to be gained the technology in developing countries. This has been proved
by applying pavement management technology in developing
countries. This has been proven in Brazil, a typical middle-income in Brazil, a typical middle-income developing country, and
developing country, and can be applied to great benefit in other can be applied to great benefit in other developing countries.
developing countries. The PMS implemented in Brazil is de- This paper summarizes the PMS implemented in Brazil for
scribed, the special limitations and standardization requirements the federal network and discusses the special limitations and
for the proper use of pavement management in a developing requirements for the proper use of pavement management in
country are discussed. Such PMS must often be done at a tech- developing countries. Such PMS must often be done at a
nology level below that in the United States. Recommendations
are presented for developing countries and for upgrading such technology level somewhat below that in large cities and states
technology. in the United States. Recommendations for standardization
are presented for developing countries and for upgrading such
technology. A brief description of PMS-related studies in
The deteriorating condition of paved road networks and the Brazil is also presented.
limited resources available for rehabilitating these roads
underscore the need for more rational approaches to select
priority links on a road network. As a result, developing and STEPS FOR IMPLEMENTING PMS
developed countries have been establishing pavement man-
agement systems (PMSs) to better manage their road infra- Experience suggests that major factors in the successful im-
structure assets. plementation and improvement of PMS are staging, preim-
A PMS consists of a comprehensive, coordinated set of plementation planning, and strong top-management support;
activities associated with the planning, design, construction, the establishment of a PMS steering committee has been use-
maintenance, evaluation, and research of pavements. Its main ful in many cases (3). An important step is that of convincing
objective is to use reliable information and decision criteria top management of the value of a PMS. Teach them what a
in an organized framework to produce a cost-effective pave- PMS can do and what is required to develop such a system.
ment program. PMS activities are directed toward achieving To this end, a formal 1-week seminar by Hudson and Haas
the best value possible for the available funds in providing was organized in Brazil in 1983; it included the participation
and operating pavements (1). of senior highway managers on the first day. A follow-up
A PMS must be able to be updated; to consider alternative seminar was given by Hudson, Haas, and Queiroz at the
strategies; to identify the optimum alternative; to base deci- University of Sao Paulo in 1985. Informal meetings, formal
sions on quantified attributes, criteria, and constraints; seminars, progress reports, and conference papers are im-
and to use feedback information about the consequences of portant means of communication between technical staff and
decisions. managers and across divisions that cooperate in PMS imple-
Considering the needs of the network as a whole, a PMS mentation. For Brazil, these activities were also crucial to the
can analyze alternative funding programs, making it possible dissemination of PMS techniques to state highway authorities .
Figure 1 shows the main stages in developing and imple-
C. Queiroz, Western Africa Department, The World Bank, 1818 H menting a PMS, which were generally followed in Brazil.
Street N.W., Washington, D.C. 20433. W. R. Hudson, Department
of Civil Engineering, University of Texas, Austin, Tex. 78712. Special constraints for implementing PMSs in developing
R. Haas, Department of Civil Engineering, University of Waterloo, countries include limited trained local human and material
Waterloo, Ontario, Canada N2L 3Gl. resources in several countries, and Brazil is no exception.
32 TRANSPORTATION RESEARCH RECORD 1344

STAGE KEY PRODUCTS/OUTPUTS

Stage 1: PRE-IMPLEMENTATION
Steering Committee approval;
Review methods and procedures; develop terms of reference for retaining
implementation plan consultant

Stage 2: DATA BASE IMPLEMENTATION

Data display formats; status of
Develop data collection procedures, data network; maintenance needs
base, output reporting

Stage 3: ANALYSIS IMPLEMENTATION

Alternative maint. & rehab.
Implement strategy analysis and program strategies; priority programs;
optimization; apply HDM and EBM models effect of varying budgets;
budget requirements for
specified standards

FIGURE 1 Major stages in development and implementation of PMS.

Adopting less-sophisticated methods and equipment for data • Optimization subsystem, which is necessary whenever the
collection can minimize the material needs. As the amount needs exceed the available funds (a common situation in de-
of modern equipment used to evaluate pavements has grown, veloping countries); and
so has the concern to choose the most-appropriate devices. • Report subsystem, which should be able to provide in-
An illustrative case is the measurement of pavement deflec- formation on the current status of the paved road network,
tion, for which the traditional Benkelman beam is of particular priority programs of rehabilitation, reconstruction and main-
interest; not only does it achieve adequate productivity at tenance, and effects of different budget levels on these pro-
generally lower costs, but its use can lead to a high degree of grams and on the state of the network.
accuracy in applying road investment analysis models such as
HDM (4). Operation of Benkelman beams is labor-intensive, For easier access, the computerized part of a PMS should
and the instruments are robust. operate on a personal computer workstation, which can even-
As for the human resource limitations, foreign consultants tually be linked to a mainframe system. A PMS should be
were used part time during the first 3 years of PMS imple- flexible in the options provided to the user and in the graphical
mentation (about 1983-86). Consultants have also been use- and tabular reporting functions ranging from detailed to
ful in PMS improvement and can help to maintain a strong summarized.
interest by agency managers through periodic visits and sem-
inars. The consultants should work closely with committed
local counterparts to insure PMS sustainability. PAVEMENT EVALUATION

A systematic approach to pavement management started in

PMS STRUCTURE Brazil in 1983 under the Brazilian National Highway De-
partment (DNER) for application on the paved federal road
The detailed structure of a PMS depends on the organization network, and several states in Brazil have gradually adopted
of the particular agency within which it is implemented. For the developed methodology.
Brazil it was considered important to include the following A specific pavement evaluation methodology has been de-
subsystems (5): veloped as part of the DNER PMS (5). For evaluation, the
paved road network is divided into homogeneous subsections.
• Information subsystem, which includes data on road length, The following sequence of procedures is applied to define
pavement type, roughness, distress, structural adequacy, traffic, sample segments, where deflections are measured, and as-
and costs. A simple and realistic procedure for periodically sessment areas, where detailed condition surveys are carried
collecting data on the road network that takes advantage of out (Figure 2):
sampling techniques was designed to best fit the needs and
capability of the federal roads agency; 1. Identification of homogeneous subsections within the un-
• Maintenance strategy subsystem, which should be able to itary sections of the National Highways Plan (PNV). These
simulate total life-cycle conditions and costs for multiple road subsections should be between 0.3 and 20 km long. The sub-
maintenance (and eventually design) alternatives for road links sections are selected visually by the resident engineer, without
making up the network. This subsystem should also assist in using any equipment. The main factors considered are the
determining current and future needs (i.e., those sections in type and condition of the surfacing and the homogeneity of
the network that have reached or will reach their minimum traffic.
acceptable or "trigger" level, depending on the criteria spec- 2. Measurement of roughness on homogeneous subsec-
ified); tions. Roughness was adopted as the primary measure of
Queiroz et al. 33

Unit Section of Less than 100 km

National Highways
Plan

General (Visual) Homogenoll.'l Subsection

Condition Survey (Variable length 0.3 to 20 km)
and Roughness
,
Deflection lOOm lOOm
Measurements .

Assessment Areas

~ ~ ~ ~
Detailed Condition 6m 6m
Survey

FIGURE 2 Sampling system for network survey.

pavement condition because it relates to safety, the overall alignment or for groups of links on an entire network. A user
economics of road transportation , and rider comfort and per- can search for the alternative with the lowest discounted total
formance. cost. If HDM is used in conjunction with the Expenditure
3. Identification in situ of sample segments considered rep- Budgeting Model (EBM), the set of design and maintenance
resentative of each homogeneous subsection. Three sample options that would minimize total discounted transport costs
segments (each 100 m long) are identified at the beginning , of an entire road network under year-to-year budget con-
middle , and end of each homogeneous subsection. straints can be determined (4). Thus the EBM model assists
4. Measurement of deflection on the sample segments. the analysis team in identifying priority sections and the best
Pavement deflection is an important parameter for predicting maintenance alternative for each priority road section . Mi-
future pavement condition . The Benkelman beam has been crocomputer versions of both HDM and EBM are now avail-
adopted to measure deflection in external wheel tracks at 20- able, which makes the models more flexible for general use.
m intervals on alternate sides of each sample segment. The models used to quantify the relationships between the
5. Survey of pavement condition. Determining the types costs of road construction and maintenance and vehicle op-
and extent of pavement defects (such as cracking, potholes, eration in HOM resulted from data collected under a collab-
and rutting) is of great importance for planning road main- orative large-scale research program, most of which was car-
tenance and rehabilitation. Six assessment areas are marked ried out in Brazil under a wide range of environmental
out on each homogeneous subsection, that is, two on each of conditions (4,6). HDM relationships are thus directly appli-
the extremes of the three 100-m sample segments. Pavement cable to Brazil and other tropical regions.
distress found in the assessment areas is duly recorded in both
qualitative and quantitative terms. OPTIMAL REHABILITATION PROGRAM

In summary, under the DNER PMS, pavement evaluation An appropriate methodology for gathering data on the road
includes a survey of pavement condition and deflection mea- infrastructure-along with data on the volume, composition,
surements on a sampling basis and of roughness measure- and weight of traffic on each homogeneous subsection-pro-
ments on the whole network. vides the basic information necessary for an econorriic analysis
Resources permitting, these measurements will continue to of alternative strategies to maintain and rehabilitate a road
be taken annually and will be summarized in a format useful network. This analysis can be applied at the project and net-
to senior management besides being used in economic and work levels using a program such as HOM.
technical analyses . The data necessary for running the HDM model refer to
the structure and condition of the existing network, the var-
ious sets of maintenance, rehabilitation and reconstruction
ECONOMIC ANALYSIS OF ALTERNATIVE alternatives, maintenance standards, unit costs, traffic pro-
MAINTENANCE STRATEGIES jections, and environmental parameters. Using these data,
the model carries out the following series of calculations:
PMS implementation requires the use of a valid model to
simulate total life-cycle performance and costs for several road •Vehicle speeds and consumption of resources;
maintenance and rehabilitation alternatives for each of the •Road deterioration and maintenance resources;
road links composing the network (4). For Brazil, the model •Road construction resources;
of choice is the Highway Design and Maintenance Standards •User, agency, and total financial and economic costs,
Model (HDM-III). calculated on the basis of physical quantities and unit costs;
HDM is designed to make comparative cost estimates and and
economic evaluations of different construction and mainte- •Net present values, internal rates of return and first-year
nance options, either for a given road project on a specific benefits.
34 TRANSPORTATION RESEARCH RECORD 1344

Running the HDM-III is divided into the following phases Benkelman beam deflections were measured) and 20,400 as-
(4): sessment areas (where detailed condition survey was per-
formed, including cracking, ravelling, pothole, patching, and
•Data entry and generation of diagnoses; rutting measurements).
• Simulation of traffic flows and variations in road condi- Running HDM to analyze 3,400 sections would be too costly
tion year by year, taking account of deterioration, mainte- and time-consuming. To make the analysis manageable, the
nance, and possible improvements; and homogeneous subsections were classified by means of a fac-
• Economic comparisons and analyses of alternative con- torial matrix with 108 cells, each of which represented a set
struction and maintenance options for selected groups of road of sections with a relatively narrow range of features. The
links. following factors and respective levels were used to define the
matrix (7):
A fourth phase is executed by using the EBM model that
selects the optimum combination of projects and maintenance •Average daily traffic (four levels): less than 1,000; 1,000
alternatives in light of budget constraints. to 3,000; 3,000 to 5,000; and more than 5,000 vpd;
An example of applying the HDM-EBM methodology is •Benkelman deflection (three levels): less than 0.5; 0.5 to
provided here by the economic analysis of alternative pro- 0.8; and more than 0.8 mm;
grams for rehabilitating the Brazilian federal paved road net- • Percentage of pavement area cracked: less than 20, 20 to
work, which took into account various annual budget levels 40, and more than 40 percent; and
for the period from 1986 to 1988. This was the first exercise •Roughness, in terms of the quarter-car index (QI): less
of this type carried out for the Brazilian federal network, and than 40, 40 to 60, and more than 60 counts/km.
it is to be followed by subsequent 3-year rolling programs.
An analysis period of 12 years was used. The main objective Ten sets of rehabilitation and maintenance alternatives were
of the analysis was to identify the sections of road that if defined for possible application to each road section (or groups
rehabilitated would result in the maximum overall net present of road sections in a cell of the factorial matrix) during the
value for each level of investment. analysis period, as given in Table 1.
Most of the benefits result from the reduction in vehicle The HDM model was then run for each cell of the factorial
operating costs over each year of the analysis period, a con- matrix, using as input data the weighted average of pave-
sequence of the improved road condition brought about by ment condition and traffic applicable to each cell. The opti-
rehabilitation or maintenance, or both. The study was rela- mum maintenance and rehabilitation alternative selected from
tively conservative in that it did not take account of the re- those shown in Table 1 was then identified for each group of
duction in accidents costs caused by safer pavements. road sections in a cell, that is the alternative that would max-
Data for the study were obtained on about 33 000 km of imize the net present value (7). Ideally, it would be desirable
the federal paved network, which resulted in about 3,400 to implement physically the optimum option for all of the
homogeneous subsections, or 10,200 sample segments (where road sections on the network. However, it was found that the

TABLE 1 PAVEMENT REHABILITATION OPTIONS

Minimum Conditions of Application

HOM Maintenance/Rehab Useful Life
(years) Roughness,
Option Code Alternative
QI Cracking
(COUOl!l/km) (%)

0 12 Routine maintenance
08 Deep Patching
12 Routine maintenance
09 Slurry seal 3 30
12 Routine maintenance
09 Double surface 30
dressing
12 Routine maintenance

4 10 Overlay (4 cm AC) 5 50
12 Routine maintenance
5 10 Overlay (8 cm AC) 5 40
12 Routine maintenance

6 10 Overlay (8 cm AC) 5 60
12 Routine maintenance
10 Overlay (12 cm AC) 5 40
12 Routine maintenance

10 Overlay (12 cm AC) 60

12 Routine maintenance
9 II Reconstruction (15 10 80
crushed stone + 5 AC)
12 Routine maintenance
Queiroz et al. 35

funds required were well over the available or plausible budget deviations for certain features of the road network. The values
levels. relating to the individual road section are as follows:
Using as input the HDM output for each cell, as well as
the most plausible budget levels, an improved version of the • Roughness, in terms of the quarter-car index in counts
EBM model was run to select priority cells to be rehabilitated. per kilometer (QI);
The criterion adopted was that of maximizing the overall net • Percentage of paved area affected by cracks rated as Class
present value. The priority sections thus identified served as 2 or worse, plus patching and potholes (CR);
the basis for a major road rehabilitation program in 1986- •Average Benkelman beam deflection in 0.01 mm (BD);
88. Actual rehabilitation design for each priority road section • Average daily volume of cars and light trucks in vehicles
was done by the application of an optimal design method (8). per day (CA);
The improved EBM model, developed at the Brazilian Road •Average daily volume of buses in vehicles per day (BU);
Research Institute, excludes any restriction with regard to •Average daily volume of trucks (medium, heavy, and
number of projects or of budgetary periods yet provides the semitrailers) in vehicles per day (TR);
same results as EBM when run with data within EBM limi- •Annual rainfall in millimt:ters per year (RA); and
tations (9). • Indicator denoting topographic relief, that is, 1 flat-
land, 2 = hilly, 3 = mountainous (RE).
ACCEPTABILITY INDEX TO PRIORITY RANK
REHABILITATION SECTIONS The values relating to the road network are expressed as
the averages and standard deviations of these eight variables.
The HDM-EBM analysis of a road network, as described, As an example, a set of values for these parameters, calculated
may become too costly and time-consuming for large road from the 1985 survey of the Brazilian federal paved road
networks. To circumvent this problem, it was deemed worth- network, is given in Table 2 (10). Calculating AI for a road
while to develop an acceptability index (AI) that could be section requires the following steps:
computed directly from the field parameters characterizing
the homogeneous subsections of a network and yet enable 1. Calculation of the reduced value-R(x)-of each of the
rehabilitation priorities to be assigned to the subsections much parameters above:
more rapidly and simply than by means of a detailed economic
analysis (10). R(x) = (x - average)/deviation
The main purpose of calculating AI is to classify the various
sections of a road network in terms of rehabilitation priorities. where
The higher the value of Al, the greater the acceptability of
R(x) = reduced value,
the section and, therefore, the smaller the need for rehabil-
x = parameter value for the section,
itation. The AI to be investigated would be allowed to range
average = average of parameter (x) for the network, and
from 0 to 100. A section with an AI of 0 would have no
deviation = standard deviation of parameter (x) for the
acceptability and would therefore be assigned maximum
network.
priority, whereas a section with an AI of 100 would receive
zero priority. The AI cannot replace the economic analysis,
but it is helpful as a means to first screen the road network 2. Calculation of the standardized value-S(x)-of each
and select a subset of road sections in most need of rehabil- of the parameters above: S(x) is obtained as the area under
itation. The HDM-EBM analysis would then be carried out the normal curve corresponding to R(x). For example, if R(QI)
on this subset and be of a much more manageable size. = 2.86, then S(QI) = 0.9979. If the standardized value is
The AI calculation algorithm was developed at the Brazilian negative, its absolute value should be used.
Road Research Institute (11). Two sets of data are necessary 3. AI can then be computed by
for calculating the Al. The first concerns the road section
itself, and the second is based on the averages and standard AI = 100 - 0.03993 x (FQI x FTR x FST) 2 ·278

TABLE 2 SUMMARY CHARACTERISTICS OF BRAZILIAN

FEDERAL PAVED ROADS
Item No. Variable Average Std . Deviation
Roughness, Qi (counts/km) 49.94 13. 86
2 Cracking (%) 18.44 26. 34
Deflection (mm) 57.38 37.99
4 Average daily traffic: cars 1,641.06 3,973 .21
5 Average daily traffic: buses 279.08 1,003.48
6 Average daily traffic: trucks 1,391.29 2,271.21
7 Rainfall (mm/yr) 1,299.45 548.08
Relief 1.90 0 .59
36 TRANSPORTATION RESEARCH RECORD 1344

where between QI (in counts per kilometer) and IRI (in meters per
kilometer) scales is (15)
FQI = [10 x S(QI)]o.o633 ;
FfR = [10 x S(CA)] 0 • 141 + [10 x S(BU)]0- 164
QI = 13 IRI
+ [10 x S(TR)] 0 •514 ; and
FST = [10 x S(CR)] 0 •038 + [10 :x S(BD))0·083
Roughness measurements for the Brazilian PMS are ob-
+ [10 x S(RA)] 0 · 024 n + [10 x S(RE)] 0 ·014 •
tained using a vehicle instrumented to produce a numeric
It is pertinent to note that the equation given to compute proportional to the vehicle response to the road traversed.
the acceptability index was developed through regression Two such systems, which are called response-type road rough-
analysis with the results of the HDM-EBM analysis of the ness measuring systems (RTRRMSs), have been used in Bra-
federal paved network (11). The objective was to obtain an zil: Maysmeters and Linear Displacement Integrator (6,16) .
Al that would approximate the priorities given by the eco- The roughness numeric obtained from a RTRRMS for a road
nomic analysis. The criterion chosen was that the Al should section depends on the test speed, type and condition of host
be correlated with the unit net present value (NPV/km) re- vehicle , and other factors. Therefore, any RTRRMS must be
sulting from the optimum rehabilitation alternative of a road periodically calibrated to produce QI or IRI.
section (JO). In fact, a comparison between priorities assigned All RTRRMSs used in Brazil are calibrated by correlation,
by the HDM-EBM analysis and those computed by the Al which is performed on control road sections. The QI (or IRI)
showed very good agreement. for each control section is obtained by a rod-and-level survey
Unit net present value (NPV/km) and the acceptability in- (17). The RTRRMS to be calibrated measures 10 or more of
dex are well correlated (coefficient of determination of 0.96) their control sections, and the results are used with QI to
by Coelho (11): determine a regression equation that is used to convert
RTRRMS measurements into QI. As a result of this calibra-
NPV/km = 1564/(AF + 1) + 164 exp( -0.098 AI) tion, roughness measurements under the Brazilian PMS are
stable over time and can be compared with careful measure-
The AI is a simple ranking approach that estimates the ments taken in any other region of the country or the world.
relative rehabilitation need (and economic return) of road
sections making up a network. It is a method that can be
applied for large road networks when the number of sections
RECOMMENDATIONS FOR OTHER
is high and the HDM-EBM optimization method may not be
DEVELOPING COUNTRIES
practical. The AI not only has a direct connection with the
optimization technique but also can be linked with the utility
Although developed for Brazil, the pavement management
concept. In a road facility management system, utility is the
technology described herein can be applied directly, or with
level of overall effectiveness that can be achieved by under-
slight adjustments, to other developing countries with large
taking a project (12). A road section with an AI of 0 has the
road networks. For countries with smaller road networks (up
highest priority to be rehabilitated, and therefore the utility
to about 50 links or homogeneous subsections), it is recom-
value of rehabilitating this section should be maximum, that
mended that the HDM-EBM optimization analysis be applied
is, equal to 100. Conversely, if a section's AI is 100, this
directly to the road links composing the network, without the
section is totally acceptable and does not require any reha-
need for grouping links into cells of a factorial matrix. Future
bilitation; the utility value of rehabilitating it should be min-
improvements to the HDM-EBM software should make it
imum, that is, zero. Therefore, the acceptability index and
feasible to apply the optimization technique directly to larger
the utility value (UV) can be linked by
networks.
The periodic computation and display of acceptability in-
UV= 100 - AI
dexes (or utility values) for a road network provides a simple
means for senior management, and eventually the public user,
to monitor performance of the networks they manage or use.
ROUGHNESS SCALE
It is recommended that an acceptability index similar to the
one shown here be computed and used in developing countries
Roughness plays an important role in pavement management
in the future.
because it is the most important measure of road condition
influencing vehicle operating costs and because it affects the
safety, comfort, and speed of travel. Roughness was expressed
in this paper in terms of QI in counts per kilometer (13). CONCLUSIONS
However, roughness measurements are now generally ex-
pressed in terms of the international roughness index (IRI) The successful development, implementation, improvement,
(14). IRI is mathematically defined from relatively true profile and sustainability of a PMS in Brazil appeared to result from
to simulate the vertical motions induced in a moving quarter- careful preimplementation planning, strong support by senior
car (i .e., one wheel, suspension , and sprung mass). management, a sound data base, use of adequate models, and
The basic concepts of QI and IRI come from simulation of a commitment by those responsible for its operation. External
vertical motion on a road profile. QI was the roughness stan- resources, in terms of specific financing and consultant ex-
dardized in the Brazil/UNDP study (6), which provided most pertise, have played a key role in all phases of PMS devel-
of the data originating the HDM model. The relationship opment and implementation .
Queiroz et al. 37

This paper described the main stages and activities in the 7. P. S. M. Coelho and C. Queiroz. Determination of Priority Sec-
development of a PMS for use in Brazil, including condition tions for Road Rehabilitation under the Brazilian Pavement Man-
agement System. Technical Report CPGP-01/85. Brazilian Road
surveys, roughness measurements, a data base for pavement- Research Institute, Rio de Janeiro, 1985.
related information, an analysis scheme (including the use of 8. C. Queiroz, J.P. Magalhaes, P. S. M. Coelho, and N. Robertson.
HDM and EBM models and an acceptability index), and im- An Optimal Design Method to Rehabilitate Low-Volume As-
plementation procedures. phaltic Roads. In Transportation Research Record 1106, TRB,
The acceptability index presented is a simple ranking ap- National Research Council, Washington, D.C., 1987, pp. 244-
251.
proach that provides the relative rehabilitation need (and eco- 9. P. S. M. Coelho. A Model to Select Year-to-Year Optimum In-
nomic return) of road sections comprising a network. It is a vestment Alternatives Under Budgetary Constraints. Brazilian Road
method that can be applied for large road networks for which Research Institute, Rio de Janeiro, 1985.
the number of sections is high and the HDM-EBM optimiza- 10. C. Queiroz, P. S. M. Coelho, and J.P. Magalhaes. Investigation
of an Acceptability Index to Classify Road Sections on a Network.
tion method may not be practical. The Al not only has a Technical Memorandum CPGP-10/85. Brazilian Road Research
direct connection with the optimization technique used but Institute, Rio de Janeiro, 1985 .
also can be linked with utility concepts. 11. P. S. M. Coelho. Development of an Algorithm to Compute an
Although developed for Brazil, the pavement management Acceptance Index to Prioritize Road Sections for Rehabilitation.
technology described in this paper can be applied directly, or Brazilian Road Research Institute, Rio de Janeiro, 1986.
12. Y. Jiang and K. Sinha . Approach to Combine Ranking and Opti-
with slight adjustments, to other developing countries. Cer- mization Techniques in Highway Project Selection. In Trans-
tainly the concepts are applicable everywhere. portation Research Record 1262, TRB, National Research Coun-
cil, Washington, D.C., 1990, pp. 155-161.
13. C. Queiroz, W. R . Hudson, A. Visser, and B. C. Butler. A
Stable, Consistent and Transferable Roughness Scale for World-
wide Standardization. In Transportation Research Record 997,
REFERENCES TRB, National Research Council, Washington, D.C., 1984.
14. M. Sayers, T. Gillespie, and C. Queiroz. The International Road
1. R. Haas and W. R. Hudson. Pavement Management Systems. Roughness Experiment: Establishing Correlation and a Calibra-
McGraw-Hill Book Company, New York, N.Y., 1978. tion Standard for Measurements. Technical Paper 45 . The World
2. W.R. Hudson. Pavement Management. Proc., 6th International Bank, Washington, D.C., 1986.
Co11fere11ce 011 the Struclural Design of Asphalt Pavements, Vol. 15. W. D. 0. Paterson . Road Deterioration and Maintenance Effects:
2, University of Michigan , Ann Arbor, 1987, pp. 97-116. Models for Planning and Management. The Highway Design and
3. L. Hill and R. Haas. Developmellt and lmpleme111ation of a Pave- Maintenance Standards Series, The Johns Hopkins University
ment Management System for Minnesota. FHWA, U.S. Depart- Press, Baltimore, Md., 1987.
ment of Transportation, 1990. 16. F. Domingues and C. Queiroz. Development of a Linear Dis-
4. T. Wa1anatada, C. Harral, W. Paterson, A. Dhareshwar, A. placement Integrator to Measure Road and Runway Roughness .
Bhandari, and K. Tsunokawa. The Highway Design and Main- Presented at the 20th Annual Meeting of the Brazilian Paving
tenance Standards Model . Johns Hopkins University Press, Bal- Association, Fortaleza, 1985 .
timore, Md., 1987. 17. C. Queiroz. Calibrating Response-Type Roughness Measure-
5. C. Queiroz, J . P. Magalhaes, and P. S. M. Coelho. The State- ment Systems Through Rod-and-Level Profiles. In Transporta-
of-the- Ari ·of Pavemenr Managemellt iJ1 Brazil. Presented at the tion Research Record 898, TRB, National Research Council,
15th Pan-American Road Congress, Mexico , 1986. Washington, D.C., 1983.
6. Research on the Interrelationships between Costs of Highway Con-
struction, Maintenance and Utilization. Final Report. Brazilian Publication of this paper sponsored by Committee on Pavement Man-
Transportation Planning Agency (GEIPOT), Brazil, 1982. agement Systems.

View publication stats