Structure and Infrastructure Engineering

Maintenance, Management, Life-Cycle Design and Performance

ISSN: 1573-2479 (Print) 1744-8980 (Online) Journal homepage: www.tandfonline.com/journals/nsie20

Methodology and base cost models to determine the total

benefits of preservation interventions on road sections in
Switzerland

B. T. Adey, T. Herrmann, K. Tsafatinos, J. Lüking, N. Schindele & R. Hajdin

To cite this article: B. T. Adey, T. Herrmann, K. Tsafatinos, J. Lüking, N. Schindele & R. Hajdin
(2012) Methodology and base cost models to determine the total benefits of preservation
interventions on road sections in Switzerland, Structure and Infrastructure Engineering, 8:7,
639-654, DOI: 10.1080/15732479.2010.491119

To link to this article: https://doi.org/10.1080/15732479.2010.491119

Published online: 23 Jun 2010.

Submit your article to this journal

Article views: 536

View related articles

Citing articles: 2 View citing articles

Full Terms & Conditions of access and use can be found at

https://www.tandfonline.com/action/journalInformation?journalCode=nsie20
Structure and Infrastructure Engineering
Vol. 8, No. 7, July 2012, 639–654

Methodology and base cost models to determine the total benefits of preservation interventions
on road sections in Switzerland
B.T. Adeya*, T. Herrmannb, K. Tsafatinosa, J. Lükingb, N. Schindeleb and R. Hajdinc
a
Institute for Construction Engineering and Management, Swiss Federal Institute of Technology, Zürich, Switzerland;
b
R þ R Burger and Partner Inc., Baden, Switzerland; cInfrastructure Management Consultants Ltd., Zürich, Switzerland
(Received 13 August 2009; final version received 30 April 2010; accepted 1 May 2010; published online 23 June 2010)

In Switzerland, it is common practice to estimate the total benefits of new roads before they have been built and of
road improvement interventions before they have been performed. The guidelines with respect to standardised
methodologies and models to be used have been developed (VSS 2003a) and are now in use. It is not yet, however,
common practice to estimate the total benefits of road preservation interventions and hence guidelines do not exist
with respect to standardised methodologies and models to be used. In order to provide this information, the research
package VSS 2004/710-716 was started. This paper presents the methodology and models proposed to be used to
evaluate the total benefits of road preservation interventions in Switzerland based on the evaluation and synthesis of
existing national and international literature on the relationships between pavement condition and benefits of road
use conducted in project 714 of the research package. The use of the methodology and models are demonstrated by
determining the optimal intervention strategy for a representative situation in Switzerland. The impact on the
optimal intervention strategy of the consideration of multiple stakeholders is investigated.
Keywords: benefit analysis; cost analysis; infrastructure management; maintenance; roads and highways

are performed at regular or irregular intervals

1. Introduction greater than 5 years, to reduce the deterioration
Road maintenance requires the planning and performing rate of road section or to restore the condition of
of interventions to ensure that an adequate level of a road section up to, but not beyond, its original
service is provided. The definition of an adequate level of condition. They are normally performed when
service is often, at least partially, defined in codes and the condition of a road section reaches a specified
guidelines. The interventions required in road main- threshold, e.g. a certain percentage of the surface
tenance that do not involve a change of the level of is cracked. Examples of preservation interven-
service, can be grouped into the following two categories: tions are asphalt crack sealing, chip sealing, and
partial and full depth concrete repairs to restore
. Routine maintenance interventions. These are functionality of the slab.
interventions performed on a routine basis, i.e.
at less than 5 year intervals, to prevent the It is in the interest of society that roads be managed
unnecessary acceleration of the deterioration to ensure that total benefits are maximised. To achieve
rate from the normally expected deterioration this goal optimal preservation intervention strategies
rate and therefore the unnecessary increase in risks must be followed. A preservation intervention strategy
related to the condition of the road section, such as consists of the types of preservation interventions and
accident risks. Routine maintenance interventions the times when they are to be performed on an
do not improve the overall condition of a road infrastructure object, in this case a road section. The
section or reduce the deterioration rate below the subsequent changes in routine maintenance interven-
normally expected deterioration rate. They are tions are normally considered as a (positive or
performed between preservation interventions. negative) benefit of the preservation intervention
Examples of routine maintenance interventions strategy. An optimal preservation intervention strategy
are cleaning of road side ditches, crack filling and is the one that has the largest positive difference
pothole filling. between its net benefit and its net cost. The net benefit
. Preservation interventions. These interventions of an intervention strategy is the difference between its
are not performed on a routine basis, i.e. they benefit and the benefit of the reference intervention

*Corresponding author. Email: [email protected]

ISSN 1573-2479 print/ISSN 1744-8980 online

Ó 2012 Taylor & Francis
http://dx.doi.org/10.1080/15732479.2010.491119
http://www.tandfonline.com
640 B.T. Adey et al.

strategy. The net cost of an intervention strategy is the benefits associated with road section preservation
difference between its cost and the cost of the reference intervention strategies and therefore to determine
intervention strategy. Costs, in this case, are the owner optimal road section preservation intervention strate-
costs of performing the intervention strategy. gies. The use of the methodology and models are
In order to provide the required methodologies, demonstrated by determining the optimal intervention
models and basic information to be used to systematically strategy for a representative road section in Switzerland.
determine optimal preservation intervention strategies The impact on the optimal intervention strategy of the
for road sections in Switzerland, the research package consideration of multiple stakeholders is investigated.
VSS 2004/710-716 was started. In addition to fulfilling
this goal, it has provided a starting point from which
future research will be conducted. The work within this 2. Benefits
research package involved exclusively the evaluation and A benefit of an intervention strategy is defined as a
synthesis of existing research work. It consisted of (positive or negative) consequence of the intervention
research projects that focused on the standardisation of strategy on a stakeholder, i.e. something that happens
interventions, the deterioration of pavement sections, the when an intervention strategy has been followed that
benefit of interventions and the additional costs of would not have happened if it had not been followed.
advancing or postponing interventions. They are grouped by stakeholder, i.e. the owner, the
This paper presents the methodology and models user and the public, and benefit type, and represented
proposed to be used to evaluate the total benefits of road with indicators (Table 1). The benefits for the owner are
preservation interventions in Switzerland based on the the reduction in routine maintenance costs. The benefits
evaluation and synthesis of existing national and for the user are the reduction in vehicle operation costs,
international literature on the relationships between the reduction in accident costs, the reduction in travel
pavement condition and benefits of road use conducted time costs and the reduction in discomfort costs. The
in project 714 of the research package. It provides the benefits for the public are the reduction in accident
main results of the work and references the principal costs, the reduction in noise costs and the reduction in
literature upon which the proposed methodology and other environmental costs. Each benefit type is attrib-
models were based. A complete list of the literature uted to one stakeholder, with the exception of benefits
investigated can be found in Herrmann et al. (2008). The due to the reduction in accidents. A reduction in
models proposed use only variables commonly used in accidents is considered to benefit both the user and the
practice to facilitate their acceptance and use in practice. general public. The benefits are orthogonal to one
As no tests were conducted, it is expected that these another and can therefore be added (or subtracted) in
models will be improved in the future when testing is the estimation of total benefits.
conducted to further investigate the relationships Total benefit is defined as the net benefit for all
between road condition and benefits. Nevertheless the stakeholders over an investigated time period [0, T], i.e.
systematic and logical methodology and models pro- the benefit for all stakeholders of one intervention
posed can be used to numerically estimate all of the strategy when compared with the benefit for all

Table 1. Stakeholders, benefit types, and benefit indicators.

Stakeholders Benefit type Symbol Example benefit indicators

Owner Reduction of routine maintenance costs (M) BiM ðtÞ Man-hours needed
Quantity of materials required
User Reduction of travel time costs (TT) BiTT ðtÞ Man-hours of work-time lost
Man-hours of non-work-time lost
Reduction of vehicle operating costs (VO) BiVO ðtÞ Man-hours needed for maintenance
Quantity of materials required, e.g. number of
tyres, number of brake pads, amount of fuel
Reduction of discomfort costs (D) BiD ðtÞ Amount of physical disturbance
Amount of psychological disturbance
Reduction of accident costs (A) BiA ðtÞ Amount of property damage
Number and type of injuries
Number of deaths
Public Reduction of environmental costs (E) BiE ðtÞ Amount of noise
Amount of air pollution

Note: i ¼ intervention strategy; t ¼ unit of time.

 Structure and Infrastructure Engineering 641

stakeholders of a reference intervention strategy In comparison to the other methodologies pre-

(Equation (1)). viously investigated or proposed, the methodology
Z T presented in this paper encompasses all of the benefits
 i and stakeholders related to the investigated road
TBi ¼ BM ðtÞ þ BiTT ðtÞ þ BiVO ðtÞ þ BiD ðtÞ þ BiA ðtÞ
0 section, how they are related to pavement condition,

þ BiE ðtÞ  egt dt ð1Þ and how they change over time. According to the
classification proposed by Khurshid et al. (2009) it
where TBi is the total benefits of intervention strategy would fall into the costs and benefits category.
I; 0 is the start of the investigated time period; T is the The proposed methodology is based on the quanti-
end of the investigated time period; and g is the fication of basic units consumed/used (e.g. number of
discount rate. The meaning of other variables is days of intervention), produced/caused (e.g. number of
explained in Table 1. accidents) or emitted (e.g. tons of CO2) on the network
When the benefits of intervention strategies are over the investigated period of time and assigning a
seen as the differences in costs, C, that would be monetary value to each of these units. Such a methodol-
incurred by following intervention strategy i and those ogy allows that determining of the monetary value of each
that would be incurred by following the reference of the benefits using utility theory (Keeney and Raiffa
intervention strategy, R, Equation (1) becomes: 1993) or using comparison techniques such as the
Z T analytical hierarchy process (Saaty 2006) and many of
 R    the willingness to pay methods, such as opportunity cost
TBi ¼ CM ðtÞ  CiM ðtÞ þ CR i
TT ðtÞ  CTT ðtÞ
0 methods and contingent valuation methods. The mone-
   R   R
þ CR i i
VO ðtÞ  CVO ðtÞ þ CD ðtÞ  CD ðtÞ þ CA ðtÞ
tary values used in this work were determined from a
   gt comparison of values found in national and international
 CiA ðtÞ þ CR i
E ðtÞ  CE ðtÞ : e dt ð2Þ
literature and were adjusted based on expert opinion for
Switzerland where necessary (Hermann et al. 2008).
The steps of the proposed methodology to evaluate the
3. Methodology total benefits of an intervention strategy are shown in
Since roads deteriorate over time due to their exposure Figure 1. The identification of the states of the road section
to the environment and to use, it is necessary to are required to take into consideration the fundamentally
maintain them to ensure that they continue to provide different system behaviour when a preservation interven-
an adequate level of service. An adequate level of tion is being performed and when no preservation
service is ensured through the timely performance of
interventions. There are many possible intervention
strategies that can provide an adequate level of service
and they all have different costs and benefits. In order
to be able to determine the optimal preservation
intervention strategy from the many different accep-
table ones it is necessary to determine both the total
costs and the total benefits of each.
In the past numerous researchers have investigated
optimal interventions strategies for pavement interven-
tions, including Lamptey et al. (2008) who studied
optimal pavement intervention strategies taking into
consideration the owner costs of intervention and the
user costs during intervention and Labi and Sinha
(2005) who studied the cost effectiveness of various
levels of pavement intervention, measuring effectiveness
as the increase in service life and taking the owner and
user costs during intervention into consideration, as
well as Al-Mansour and Sinha (1994), Geoffroy (1996)
and O’Brien (1989). In 2009 a comparison of methods
for evaluating pavement interventions was conducted
(Khurshid et al. 2009) which, among other things, found
that the criteria used in determining optimal interven-
tion strategies could be grouped as methods measuring Figure 1. Steps to evaluate the total benefit of an
effectiveness, costs only, or costs and benefits. intervention strategy.
642 B.T. Adey et al.

intervention is being performed. Two states are often incurred during preservation interventions. The equa-
adequate, but for complex preservation interventions that tions proposed to be used to estimate maintenance
last a relatively long period of time it may be required to costs, CM, between (q ¼ 1) and during preservation
have more than two states. interventions (q ¼ 2) are:
The road condition in Switzerland is approximated 
based on pavement condition, which is measured using CRM  CFRM ðI1 Þ  s; q ¼ 1
CMq ¼ ð3Þ
six indices (VSS 2003b). I0 is used to measure surface Cp ; q¼2
damage without taking ruts into consideration. I1 is where CRM is the routine maintenance costs when
used to measure surface damage taking ruts into I1 ¼ 3; CFRM ðI1 Þ is a routine maintenance cost factor
consideration. I2 is used to measure longitudinal which varies as a function of I1 as:
unevenness. I3 is used to measure transversal uneven- 8
ness. I4 is used to measure surface friction. I5 is used to >
< 1= for I1 2 ½0; 1Þ
3
measure the load carrying capacity of the road section. F
CRM ðI1 Þ ¼ 1=3  I1 for I1 2 ½1; 3Þ
>
: 25 2 73 75
=12  I1  =6  I1 þ =4 for I1 2 ½3; 5Þ
4. Cost models ð4Þ
In general it is proposed to approximate costs by Due to limited available relevant data, the relation-
estimating the relevant units, such as road section ship between the routine maintenance costs and
length upon which each vehicle travels (s), daily traffic pavement condition was determined based principally
volume (DTV), travel time per vehicle (t), and number on the experience and opinion of a number of experts
of people affected (p), and multiplying them by the in Switzerland who are routinely involved in making
appropriate unit costs. Two sets of basic cost models predictions of maintenance costs. It was determined as
(q) are proposed in this section. One set is to be used follows:
between preservation interventions, i.e. when no
preservation intervention is being performed (q ¼ 1) . The mean costs CRM as reported in Müller (2005)
and the other is to be used during preservation were assumed to be associated with the mean
interventions, i.e. when a preservation intervention is condition of cantonal roads in Switzerland
being performed (q ¼ 2). All important roads onto I1 ¼ 3 (Bürgi 2002, Hufschmid 2002).
which vehicles may be deviated during an intervention . The maintenance costs when the pavement is in
are to be taken into consideration, as well as the excellent condition (0  I1 5 1) were set at one
percentage of vehicles deviated on to them, in the third of the mean based on cantonal experience
estimation of all benefits. The elasticity of traffic in Switzerland (Goebbels et al. 2008) taking into
demand should also be taken into consideration. consideration, for example, that there will be
Example suggested unit costs to be used in Switzerland some damage that must be repaired due to
for each model are given in each section. These are accidents or plowing in winter that occurs
based on values found in national and international regardless of road condition.
literature and adjusted using expert opinion. Details . It was then assumed in the absence of informa-
with respect to their determination can be found in tion in literature that the routine maintenance
Herrmann et al. (2008). All costs are given in 2007 costs for fairly good pavement conditions in-
Swiss Francs. crease linearly between (1  I1 5 3).
. The maintenance costs when the pavement is in
very poor condition (I1 ¼ 5) were set to be 10
4.1. Maintenance costs times higher than for roads in sufficient condition
Maintenance costs are defined as all costs to the owner (I1 ¼ 3), based on the work of Hufschmid
of performing interventions, e.g. the cost of labour and (2002).
material, that prevent or impede deterioration or . The quadratic increase in maintenance costs
improve the road up to, but not beyond, its original between when the pavement condition is 3 
condition. Maintenance costs are considered to be I1  5, was established based on the non-
composed of routine maintenance costs (CRM) and linearity in maintenance costs over this range
preservation costs (Cp) (Equation (3)). Routine main- reported by Hufschmid (2002). This non-linear
tenance costs, which are defined in the introduction, increase is meant to take into consideration the
are assumed to be incurred between, but not during, non-linear increase in routine maintenance inter-
preservation interventions and vary directly as a ventions, such as the repairing of pot holes,
function of road condition. Preservation costs, which which are required simply to maintain the
are also defined in the introduction, are assumed to be minimal acceptable service level, over this range.
 Structure and Infrastructure Engineering 643

where g is the percentage of daily traffic that occurs

4.2. Travel time costs during the analysed time unit; vtech is the free flow
Travel time costs are the costs of travelling in terms of traffic speed, which can be calculated as
time lost. There is a distinction made between lost time
that would have been used for a productive activity nqtech ¼ min ðnSignal ;q nqRough Þ; ð8Þ
and lost time that would have not been used for a
productive activity. They are assumed to be carried by where vSignal is the posted speed limit, i.e. it is assumed
the user. Some of the difficulties involved in determin- that the driver travels at the maximum posted speed
ing the relationship between pavement condition and limit if there is no congestion; and vRough is the
travel time costs are: adjustment of the free flow speed, which is dependent
on the pavement condition and calculated as:
. The relatively few tests that have isolated the
road condition in the evaluation of its impact on vSignal  a I2 b when I2  b
vRough ¼ c ð9Þ
vehicle speed. vSignal when I2 < b
. The continuous variation of speed along a road
section, due to multiple factors, such as traffic where a is the speed penalty per IRI unit1 which is
behaviour, making it relatively difficult to obtain dependent on road and vehicle type (2.3 suggested); b
isolated information about speed and condition is the speed reduction starting when IRI is higher than
at exactly the same location. 5 m/km (Table 2) (Bennett and Greenwood 2000); c is
. The dependency of vehicle speed on road geome- the factor which gives the degree with which an I2-
try, making comparisons between the studies Index point per IRI-unit is to be penalised (Table 2)
published in literature virtually impossible. (Bennett and Greenwood 2000)2.
The unit costs suggested for freight transport are
Acknowledging these difficulties, the base model 15.60 CHF/vehicle-hour for the vehicle and 19.70 CHF/
proposed for the estimation of travel time costs is person multiplied by the number of passengers per vehi-
shown in Equation (5). The costs vary indirectly as a cle (VSS 2007). The unit costs suggested for passenger
function of pavement condition, i.e. pavement condi- traffic are 19.70 CHF/person, multiplied by the average
tion affects the speed of driving, and therefore, the number of passengers per vehicle (VSS 2003c). Assuming
amount of time required to complete a specific trip, an occupancy rate of 1.57 persons per car, the unit costs
once certain road condition thresholds have been equal 30.93 CHF/vehicle-hour for passenger transport.
reached. These thresholds are defined per road
category. The relationship between speed and pave-
ment condition was principally determined based on 4.3. Vehicle costs
the results in (Bennett and Greenwood 2001); one of Vehicle costs are the costs of maintenance and operation
the most complete studies in this area. I2 was of a vehicle. The costs of vehicle operation are principally
considered to be an adequate representation of comprised of fuel costs. The costs of vehicle maintenance
pavement condition for the estimation of travel time include the cost of the man-hours needed for maintenance
costs. The travel time costs CTT are calculated as: work and the cost of the materials required, e.g. tyres and
brake pads. Some of the difficulties involved in determin-
CTTq ¼ UCTTq  tq  DTVq ; q ¼ 1; 2 ð5Þ ing vehicle costs are that:
. vehicles are driven over many different roads in
where UCTT is the unit cost of travel time per vehicle
many different conditions, making it difficult to
  ! link costs to road condition,
s uq b
tq ¼ q 1þa ð6Þ . vehicle costs depend significantly on driver
vtech capq
behaviour, which in turn depends on the traffic
volume and behaviour,
where a is the parameter dependent on road character-
istics (0.15 suggested; Lee and Machemehl 2005); b is
the parameter dependent on road characteristics (4.0 Table 2. Suggested values of b and c as a function of road
suggested; Lee and Machemehl 2005); cap is the road type.
capacity, expressed as number of vehicles per unit Road type b c
time; u is the traffic flow during analysed interval,
calculated as Highway 4.2 0.6
Main rural road 3.7 0.7
Secondary rural road/city road 3.2 0.6
u ¼ g  DTV ð7Þ
644 B.T. Adey et al.

. vehicle costs depend significantly on road geo- 4.4. Discomfort costs

metry, including the slope, curvature, and road Discomfort costs are the costs of variation in comfort of a
width, and travelling individual, both physical, e.g. bruises from an
. vehicle efficiency changes over time and varies extremely bumpy ride, and psychological, e.g. anxiety due
between vehicles, making it difficult to compare to a perceived increase in the probability of having an
the results of studies conducted in different accident. They are considered to be carried by the user. The
countries over different periods of time. principal difficulty in the estimation of discomfort costs is
the high subjectivity of the value of discomfort and the lack
In the base model proposed, vehicle costs are of objective measures to determine whether it occurs. For
grouped as fixed costs, i.e. costs that can be approxi- example, the same person could drive over a rough road
mated per time unit travelled, and variable costs, i.e. section today and be very unhappy about the roughness,
costs that can be approximated per travelled distance and tomorrow, with a lot on his mind, not even notice it.
unit, or per unit of fuel consumed. The fixed costs are Although various researchers have shown that
indirectly dependent on road condition, since road there is a link between discomfort and the longitudinal
condition affects the speed of travel and, therefore, unevenness of pavement (e.g. Ihs 2005, LTNZ 2006),
travel time. The variable costs are directly dependent on there is little indication that there is any variation in
pavement condition, principally through the longitudi- driving comfort when the pavement condition is
nal unevenness of the road (LTNZ 2006) and (Pichler between very good and acceptable (0  I2  3), which
1981). Combining the fixed costs and the formula for is normally the case in Switzerland, and the values
estimation of the variable costs as proposed by determined for worse pavement condition (3  I2  5)
(Pichler 1981) the vehicle costs can be modelled as in are widely disputed among experts. It is therefore
Equation (10). proposed to estimate discomfort costs CD, both
between and during intervention strategies, as:
  
13I2 787 CDq ¼ 0; q ¼ 1; 2 ð11Þ
CVq ¼ DTVq  tq  UCVH þ þ
800 800

4.5. Accident costs
 ðsq  ðUCVkm þ Tvq  UCf ÞÞ ; q ¼ 1; 2 ð10Þ
Accident costs are the costs that result from having
an accident, including property damage, injuries and
where CV is the vehicle costs; UCVH is the unit vehicle deaths. They are estimated using the accident rate and
costs per hour driven; UCVkm is the unit vehicle costs the cost of each accident when it occurs. The main
per kilometre driven; Tv is the fuel used per kilometre difficulties in the linking of accident costs to pave-
at the driven speed; UCF is the unit cost of fuel; ment condition is that the majority of accidents
  depend principally on driver mistakes (Cleveland
13I2 787 1987, Henry 2000, Bester 2003, Stütze 2004, Noyce
þ
800 800

is an empirically derived factor to take into considera-

tion that the vehicle costs depend on the distance
travelled and that fuel consumption over this distance
depends on the longitudinal unevenness of the road
(Pichler 1981). The value of this factor is between 0.984
and 1.065 (Pichler 1981).
The suggested cost values are given in Table 3
(ASTRA 2003). The relationship between travel speed
and fuel consumption is shown in Figure 2 (Keller and
Zbinden 2004).

Table 3. Suggested unit costs for the determination of

vehicle costs.

Unit cost Unit Passenger vehicle Truck

UCVH [CHF/h] 1.91 6.16
UCVkm [CHF/km] 0.19 0.44
UCF [CHF/l] 0.53 0.58 Figure 2. Suggested values of fuel use as a function of
speed.
 Structure and Infrastructure Engineering 645

et al. 2005) and pavement condition only has an where y(I4) is the rate of occurrence of specific
impact on their ability to correct these mistakes once damage type dependent on road condition when no
they are made, for example through shortening the intervention is being performed. It is approximated
breaking distance. It is also complicated by the fact using the I4 measure of surface friction and road
that the increases in friction and decreases in type (suggested values for highways are given by and
unevenness sometimes result in higher accident rates shown in Equations (13) to (15) (Stütze 2004)),
(Cleveland 1987, Hauer 1997, Schandersson 1994, suggested values for inside and outside city limits are
McLean 1995, Bester 2003). given in Table 4),

67
ypropertyq ðI4 Þ ¼ 0:4: ð13Þ
99:8541 þ 46:572  lnð0:00981  I24  0:1081  I4 þ 0:5966Þ

67
yinjuriesq ðI4 Þ ¼ 0:16: ð14Þ
99:8541 þ 46:572  lnð0:00981  I24  0:1081  I4 þ 0:5966Þ

67
ydeathsq ðI4 Þ ¼ 0:0030: ð15Þ
99:8541 þ 46:572  lnð0:00981  I24  0:1081  I4 þ 0:5966Þ

Taking into consideration the multitude of where i is the type of damage, either property
research in this area, it is proposed to estimate damage, injuries and deaths; UCi is the unit costs of
accident costs as shown in Equation (12), where they damage (suggested values are given in Table 5
are assumed to vary directly as a function of road (ASTRA 2003); and c is a correction factor to
condition. The relationship between the accident rate take into consideration the effect of the work site
and road condition was determined based principally while performing an intervention (q ¼ 2). For q ¼ 1
on the results in (NRPA 1997, ASTRA 2003, Ihs c equals 1.
2004, Keller and Zbinden 2004, Stütze 2004). It is
considered that accident costs are partially carried by
the user (e.g. the value of a temporary or permanent 4.6. Environmental costs
disability to the user, often measured as willingness Environmental costs are considered to be made up of
to pay to avoid injury) and partially carried by noise costs, CN, air pollution costs, CAP, and climate
society (e.g. the value of a temporary or permanent costs, CC, and are calculated as:
disability to society, often measure in terms of lost
economic output). The accident costs CA are cal- CE ¼ CN þ CAP þ CC : ð16Þ
culated as:
4.6.1. Noise costs
X
I Noise costs are costs due to the disturbance of persons
CAq ¼ ðyiq ðI4 Þ  UCi  DTVq  sq  cÞ; q ¼ 1; 2 through excessive noise. Examples of this disturbance
i¼1 are headaches, high blood pressure and sleep problems.
ð12Þ In extreme cases the disturbance could even be

Table 4. Suggested values of y(I4).

y(I4) per 1 Mio. vehicle-km

Inside city limits Outside city limits
I4 Property Injuries Deaths Property Injuries Deaths
0.0–1.5 2.20 0.92 0.0098 0.70 0.4100 0.0138
1.5–3.0 2.75 1.15 0.0123 1.05 0.5125 0.0173
3.0–4.0 5.50 2.30 0.0246 2.10 1.0250 0.0346
4.0–5.0 8.80 3.68 0.0392 2.80 1.6400 0.0552
646 B.T. Adey et al.

Table 5. Suggested values of unit costs of damage.

User cost Public cost

Damage type Symbol Suggested amount CHF User cost CHF CHF
Property damage UC1 45,100 70% 31,570 13,530
Injuries UC2 293,500 80% 234,800 58,700
Deaths UC3 3,645,000 85% 3,098,250 546,750

psychological illness. They are mainly composed of Table 6. Suggested values of the correction factors pave-
rolling noise and motor noise. Some of the difficulties ment types and slopes
involved in the estimation of noise costs include the fact Pavement type DLB Slope DLs
that they are not directly quantifiable. Their quantifica-
tion requires the consideration of such things as: Melted asphalt þ0 g%, where g 4 0 0.8 g
Concrete þ3 g%, where g 5 0 0
Pavement þ8
. The variation of rental prices, which vary due to Drained asphalt 75
many different and difficult to isolate factors.
. The noise attributed health costs, which many
also occur due to many different and difficult to
isolate factors. is the motor noise for vehicle type j, for passenger
vehicles calculated as:
It is assumed that noise costs are borne by the   v 3:5 
public and it is proposed that they be estimated as LPCA ¼ 62:7 þ 10 log 1 þ ð20Þ
shown in Equation (17). Since the main factors 44
affecting noise are the speed of the vehicles and the and for trucks calculated as:
type of pavement, they are assumed not to vary   v 3:5 
directly as a function of road condition. LTA ¼ 76:9 þ 10 log 1 þ ð21Þ
56
CNq ¼ pq  UCdBAq  Leqq ; q ¼ 1; 2 ð17Þ LjR is the roll noise for vehicle type j, for passenger
vehicles calculated as:
where pq is the number of affected persons, i.e. the
number of persons that live and/or work near the road LPCR ¼ 9:5 þ 35 logðvÞ; ð22Þ
section and experience high noise levels (e.g. over
55dB(A) during day-time and over 45 dB(A) during and for trucks calculated as:
night-time); Leq is the equivalent continuous sound level
for a road section, with respect to traffic generated sound LTR ¼ 18:5 þ 35 logðvÞ; ð23Þ
events, which is calculated as given in (EMPA, 1997):
! where S is the slope category; v is the speed; UCdBA is
1X2
0:1Lj the cost per decibel, per person.
Leq ¼ 10 log ðFzj 10 Þ ; ð18Þ
K j¼1 It is noted that the cost per dB(A) per person over a
threshold value of 55 dB(A) is set to 115 CHF/Pers/
where Fzj is the number of each vehicle type/hour; K is dB(A). This value corresponds to a cost of 15.80 CHF/
the capacity of the road (vehicles/hour); j is the 1000 vehicle-km (ASTRA 2003, ECOPLAN et al.
indicator of vehicle type, either passenger car (PC) or 2004). This can also be expressed by vehicle type, i.e.
truck (T); Lj is the sound level per vehicle type, and is KdBAPC ¼ 109 CHF for passenger vehicles and
calculated as given in (EMPA 1997): KdBAT ¼ 165 CHF for trucks.
 LjA þDLS LjR þDLB 
Lj ¼ 10 log 10 10 þ 10 10 ð19Þ
4.6.2. Air pollution costs
Air pollution costs are those that are related to the
where B is the pavement type; DLB is the change in roll damage caused directly by polluted air, such as health
noise per pavement type (suggested values given in problems and the impact on plant growth. Air
Table 6); DLS (L#S is the change in motor noise per pollution is caused by the emissions of vehicles
slope category (suggested values given in Table 6); LjA travelling on the road network and in the equipment
 Structure and Infrastructure Engineering 647

required to perform preservation interventions, when where Cp–o is the owner cost of the preservation inter-
they burn fossil fuels. These emissions include carbon, vention (CHF); UAPCp–o is the unit of air pollution per
nitrogen, water, carbon monoxide, aldehydes, nitrogen owner cost of preservation intervention (ton/CHF) (sug-
dioxide, sulphur dioxide and polycyclic aromatic gested values are given in Table 7 (ECOPLAN 2007));
hydrocarbons. It is assumed that air pollution costs UCEPM10 is the unit cost of emissions (CHF/ton).
are borne by the public. Some of the difficulties The suggested value for UCEPM10 , is 28,675 CHF/
involved in estimating air pollution costs include: ton. The suggested values for UCPC T
PM10 and UCPM10
are, 0.0231 CHF/vehicle-km, 0.1474 CHF/vehicle-km,
. The determination of the number of people and respectively (ECOPLAN 2006, ECOPLAN 2007). The
plants affected by the air pollution, as the suggested values for worksite emissions, estimated as a
transport of air pollution is heavily dependent percentage of the owner cost of the intervention are
on the global weather system which results in given in Table 7.
wide variations in the distribution of air pollu-
tion on a daily if not hourly basis.
. The impact of the air pollution on human health 4.6.3. Climate costs
and plant growth, as the way living organisms Climate costs are those that are related to the damage
react to changes in emission concentrations caused indirectly by polluted air, i.e. through the chang-
varies significantly from human to human and ing of the climate which in turn may adversely affect
plant to plant. humanity. Climate change is expected to occur predomi-
nantly through the emission of CO2 into the atmosphere.
Despite being an imperfect representation the CO2 is emitted into the atmosphere through the burning
amount of PM10 emissions is used to take into of fossil fuels. Some of the difficulties involved in
consideration the many different emissions included estimating climate costs include the modelling of:
in air pollution. The air pollution costs, CAP, are
modelled as consisting of two parts as shown in: . the impact of emissions on the environment, and
. the evaluation of the severity of this impact on
CAPq ¼ CFZ E
PM10q þ CPM10q ; q ¼ 1; 2 ð24Þ humans.

where CFZPM10q is the costs of the PM10 emissions by The simple model proposed to estimate climate costs is
vehicles travelling on the road network, which are based on fuel consumption, which is affected by pavement
taken as dependent on fuel usage and therefore varies condition through its impact on vehicle speed (Section
directly as a function of the speed driven and hence 4.2). Any negative impact on the climate of performing an
indirectly as a function of road condition, measured intervention is then taken into consideration through the
using the indicator for longitudinal unevenness. They additional fuel consumption while performing the inter-
are calculated as: ventions, for example from the stopping and starting of
! vehicles. If there is reduction in fuel consumption while
XW w
FZ w UCPM10 performing an intervention, for example due to the
CPM10q þ sq  DTVq  mwq  Tv  ; q ¼ 1; 2
w¼1
Twavg reduction of speed of vehicles, there will be a climate
benefit. It is assumed that climate costs are carried by the
ð25Þ
public. The climate costs CC are calculated as
where mW is the percentage of vehicles of each type (w); !
X
W
Twavg is the fuel used on average by vehicle type (l/ CC ¼ Sq  DTVq  w
mwq  Tvq  gwq  UCCO2 ;
100 km) (suggested values are 29.1 l/100 km for w¼1
trucks, 7.7 l/100 km for passenger vehicles with diesel
q ¼ 1; 2 ð27Þ
motors, and 8.6 l/100 km for vehicles with gasoline
motors); Twv is the fuel usage per vehicle type (litres/
Table 7. Suggested worksite emissions for the evaluation of
100 km), which is dependent on the speed, v; UCwPM10 is interventions.
the average unit cost of air pollution per vehicle type
(CHF/vehicle-km); and CEPM10q is the costs of PM10 Tons PM10 per Mio.
emissions by equipment used during the performing of Emission source CHF of intervention
preservation interventions, which are taken as depen- Engines of equipment used 0.085
dent on the intervention on the worksite
 Dust from the worksite 0.104
0 q¼1 Transport to and from the worksite 0.018
CEPM10q ¼ ð26Þ Total 0.207
Cpo  UAPCpo  UCEPM10 q ¼ 2
648 B.T. Adey et al.

where UCCO2 is the unit cost of CO2 (CHF/ton) deterioration continues at 0.12 units per year. A
(suggested as 50 CHF/ton); mw is the percentage of discount rate of 2% is used.
vehicles of each vehicle type; TWv is the fuel consump-
tion of a vehicle of vehicle type W, which is dependent
on speed (l/100 km); gw is the ratio of kg CO2 per litre
fuel determined per vehicle type (suggested as 5.1. Optimisation
2.647 kg/l for diesel vehicles and 2.404 kg/l for gaso- The general mathematical formulation of the optimi-
line vehicles). sation problem is given in Equation (28).
X
A  
5. Example Max: TBMax
a  ya ð28Þ
a¼1
To illustrate how the proposed methodology and
base costs models can be used to evaluate interven- where a is the number of interventions included in
tions and to determine optimal intervention strate- intervention strategy; A is the maximum number of
gies an example was done for a representative road interventions to be considered; TBMax
a is the maximum
section in Switzerland. The example problem is to total benefit if only a interventions are performed; and
determine the intervention strategy that maximises ya is a binary variable.
total benefit over the next 75 years for a road section
where road users have two possible detours when an
intervention is being performed. The basic informa- 5.2. Analysis
tion with respect to each is shown in Table 8. Using the equations presented in the article and
It has been predetermined that the intervention exhaustive enumeration, the optimal intervention
strategy selected should consist only of partial depth strategy was determined to be to perform interven-
resurfacing interventions with reinforcement that will tions, i.e. a partial depth resurfacing with reinforce-
use cold milling in the removal of approximately ment in year 26 and 51 for a total benefit of 9.92
85 mm of the surface layer. Each intervention, when million CHF (Table 9, indicated as Intervention
an intervention is performed, will last 4 months and is Strategy 2 Int). This strategy will result in additional
performed between the 1st of May and the 31st of costs for the owner of 0.19 million CHF and reduced
August in the year it is executed. The intervention will costs for the user and for the public of 9.09 million
improve the deterioration indices of I1 to 0, I2 to 0.5 CHF and 1.02 million CHF, respectively. Table 9
and I4 to 0.5 regardless of the values of the deteriora- shows the absolute costs estimated for the reference (0
tion indices at the time of intervention (Gnehm 2008), Int.) and the optimal intervention strategy (2 Int.), and
but does not alter the deterioration rate, i.e. the benefit of the optimal intervention strategy, i.e. the

Table 8. Example: Basic information.

Variable Investigated road section Detour 1 Detour 2

Length (m) 1006 6700 11,000
Width (m) 10.5 N/A N/A
Number of lanes 2 4 2
Speed limit (km/h) 70 120 kph 50 kph
I values Decreases from initial 2 2
values at 0.12 units per year
Average daily traffic in 2010 18,700 * *
Growth in average daily traffic per year 2% of 2010 ADT 2% of 2010 ADT 2% of 2010 ADT
Traffic capacity/day 50,000 ** **
Percentage of passenger vehicles 80 N/A N/A
Percentage of trucks 20 N/A N/A
Routine maintenance 3500 N/A N/A
cost/km/year (CHF)
Persons affected by excessive 100 50 50
noise levels in 2010
Growth in persons affected by noise 2 people per year 1 person per year 1 person per year
(2% of 2010 value) (2% of 2010 value) (2% of 2010 value)
Traffic deviated during intervention 10% of vehicles 1% of deviated vehicles 9% of deviated vehicles
choose to take detours

Notes: *Assumed to be sufficiently low with respect to traffic capacity that free flow traffic exists even with the addition of detoured vehicles.
**Assumed to sufficiently high with respect to the traffic flow with detoured vehicles to allow free flow traffic to occur.
 Structure and Infrastructure Engineering 649

Table 9. Example: Absolute cost and benefits of reference and optimal intervention strategies.

Absolute costs/benefit (6 106 CHFs)

Owner User Public
Intervention Year(s) of Absolute cost Travel
Strategy Type Intervention or benefit Maintenance Time Vehicle Accident Environmental TOTAL
0 Int. 0 Cost/cost type 0.53 168 133 37 44 383
Cost/stakeholder 0.53 330 52
2 Int. 26, 51 Cost/cost type 0.72 166 130 32 44 373
Cost/stakeholder 0.72 321 49
Benefit/cost type 70.19 2.14 2.54 5.66 70.23 9.92
Benefit/stakeholder 70.19 9.09 1.02

Figure 3. Example: Expected values of the deterioration

index I1.

absolute costs when the optimal intervention strategy

is followed minus the absolute costs when the reference
intervention strategy is followed. The expected values
of the deterioration indices if the optimal intervention
strategy or the reference intervention strategy is
followed are shown in Figure 3. The deterioration
indices vary between 0.0 and 3.0 when the optimal
intervention strategy is followed. The yearly benefits
and their distribution by cost type and by stakeholder
are shown in Figure 4.
It can be seen from Table 9, that most of the costs
related to the investigated road section are incurred by
the user (330 and 321 6 106 CHF for the reference Figure 4. Example: Discounted yearly benefits and their dis-
and optimal intervention strategy, respectively). The tribution when the optimal intervention strategy is followed.
costs incurred by the public are only 15% of these costs
and those incurred by the owner are only 2% of these
costs. As most of the costs are incurred by the user so the optimal intervention strategy is followed rather
are most of the benefits that result from the choice than the reference strategy. Once I1 4 ¼ 4 the further
between intervention strategies (Figure 4). Most of the deterioration of the road condition that would occur
user benefits when the road section has an I1 5 4 are when following the reference strategy leads to increas-
due to the reduction in vehicle and accident costs when ing benefits due to the reduction in travel time costs.
650 B.T. Adey et al.

This occurs because with I4 values higher than or equal determination of optimal intervention strategies, the
to 4, as in the reference strategy, vehicles must reduce optimal intervention strategies were determined for
their speed significantly from the posted limit, and two additional cases, taking into consideration
therefore any improvement of I4 results in benefits due
to the reduction of travel time costs. . only owner benefits (Case 1), and
The performance of interventions results in a . only owner and user benefits (Case 2).
substantial increase in costs for all stakeholders in
the years of intervention and in an increase in the costs These cases along with the original, i.e. total benefits
of almost all cost types (Figure 4). The one exception is are presented together in Table 10.
the reduction in accident costs during the year of the From Table 10 it can be seen that if only owner
second intervention. This occurs because there are benefits are considered, i.e. maintenance benefits, the
more accidents when the road section has an I1 value optimal intervention strategy is to perform one
of 5 for 12 months (the reference strategy) than when intervention in year 39 resulting in 0.07 6 106 CHF
the road section has an I1 value of 2.9 for 4 months, owner benefit and total benefit of 6.96 6 106 CHF
has an intervention for four months in which the (indicated as 1 Int. in Figure 6). The costs of all
number of accidents increases, and has an I1 value of 0 intervention strategies with two or more interventions
for 4 months (the optimal intervention strategy). The are more expensive for the owner than performing one
exact distribution of costs and costs per stakeholder intervention in year 39. In fact, they would cost the
during each intervention included in the optimal owner more than even the reference strategy of doing
intervention strategy are shown in Figure 5. no preservation interventions and continuing to only
perform routine maintenance, even taking into con-
sideration the exponential relation of routine main-
5.3. Stakeholder significance tenance costs that is expected (Figure 7). The
To investigate the significance of taking into consid- maintenance costs due to the reference strategy
eration the total benefits related to a road section in the increase to 15 6 103 CHF/year but then stay constant

Figure 5. Example: Discounted benefit during interventions when following the optimal intervention strategy.
 Structure and Infrastructure Engineering 651

Table 10. Example: Absolute cost and benefits of the optimal intervention strategies for each stakeholder.

Absolute costs/benefit (6106 CHFs)

Owner User Public
Stakeholder Year(s) of Absolute cost Travel
considered Intervention or benefit Maintenance Time Vehicle Accident Environmental TOTAL
Owner 39 Benefit/cost type 0.07 2.18 1.63 3.20 70.11 6.96
Benefit/stakeholder 0.07 6.30 0.59
Owner, user 26, 51 Benefit/cost type 70.19 2.14 2.54 5.66 70.23 9.92
Benefit/stakeholder 70.19 9.09 1.02
Owner, user, 26, 51 Benefit/cost type 70.19 2.14 2.54 5.66 70.23
public
Benefit/stakeholder 70.19 9.09 1.02 9.92

Figure 6. Example: Discounted yearly benefits by cost type and stakeholder when the owner’s optimal intervention strategy is
followed.
652 B.T. Adey et al.

Figure 7. Example: Discounted owner costs when the reference intervention strategy is followed.

Figure 8. Example: Discounted benefits during intervention of the owner’s optimal intervention strategy.

(without discounting). The benefits of executing one It is interesting to note that, in this example, the
intervention in year 39 are shown in Figure 8 consideration of user benefits alone is sufficient to
(indicated as 1 Int.). determine the optimal intervention strategy. Of course
If only owner and user benefits (or only user this is because the most significant portion of absolute
benefits) are considered, the optimal strategy is the costs related to the road section is incurred by the
same as when the benefits of all stakeholders are taken user (approximately 85%) and the largest possible
into consideration, i.e. to execute interventions in years reduction in costs is due to the reduction of user
26 and 51 (650 6 106 CHF), resulting in a total benefit costs. Therefore, a 1% reduction in user costs from
of 9.92 6 106 CHF. The reason the optimal interven- the reference strategy (3.3 million CHF) is more
tion strategy, and therefore the benefits for each than six times all the owner costs of the reference
stakeholder, are the same as when the benefits of all strategy.
stakeholders are considered is due to the large portion
of total benefits that are attributable to the user
(approximately 90%). In other words, once the 6. Conclusion
optimal intervention strategy for the user is found, The methodology and models proposed to be used to
the impact of the consideration of the other stake- evaluate the total benefits of road preservation
holders has little impact on the optimality of the interventions in Switzerland were presented, along
intervention strategy. This strategy would also be with an example of how these relationships can used
approximately 120 6 103 CHF more expensive for the to determine optimal intervention strategies taking
owner than the optimal strategy determined consider- into consideration the total benefit related to a road
ing only owner costs. section.
 Structure and Infrastructure Engineering 653

The models relate benefits to road condition through ECOPLAN, 2007. External costs in road transportation.
pavement condition indicators, such as surface friction, Switzerland: Swiss Federal Roads Authority (ASTRA)
(in German).
surface damage and the longitudinal unevenness, and ECOPLAN, Planteam GHS AG and IHA-ETH, 2004.
road type. They have been developed using the natural Update for the year 2000. Noise costs of road and rail
units of the physical quantities involved, which allows transportation in Switzerland. Switzerland: Swiss Federal
for the changing of their monetary value without Authorities for Spatial Planning (ARE), Health (BAG),
modification to the methodology and models. Using and the Environment (BUWAL).
EMPA, 1997. Report on the F þ E-project ‘New EMPA model
the pavement condition information with commonly for road noise’. Switzerland: Swiss Federal Laboratories for
available information about the road section, such as Materials Testing and Research (EMPA) (in German).
daily traffic volume, speed limits or number of affected Geoffroy, D.N., 1996. Cost-effective preventive pavement
persons, the total benefit can be numerically estimated maintenance. National Cooperative Highway Research
and attributed to each stakeholder group. Program Synthesis of Highway Practice 223. United
States of America: Transportation Research Board.
To ensure that the presented methodology and Gnehm, 2007. Standard interventions for road maintenance.
models can be used to accurately evaluate road preser- VSS 2004/711. Switzerland: Swiss Federal Roads
vation interventions in Switzerland, it is now necessary Authority (in German).
to determine the most appropriate values of the phy- Goebbels, D., Albrecht, J., Rafi, A., and Bär, H., 2008. Road
sical quantities to be used in the models and to compare condition and maintenance costs. Deliberations of the VSS
Expert Group 7.02 on costs, benefits and values. Switzer-
predicted results with real world examples. It is also land: The Swiss Association of Road and Transportation
advised that research work be conducted to verify and Experts.
improve upon the base cost models if necessary. Hauer, E., 1997. Observational before-after studies in road
safety: Estimating the effect of highway and traffic
measures on road safety. Oxford, England: Pergamon
Notes Press/Elsevier Science Ltd.
1. IRI is the international roughness index. It is a world- Henry, J., 2000. Evaluation of pavement friction characteristics:
wide standard for measuring pavement roughness. The A synthesis of highway practice. NCHRP 291. United
index measures pavement roughness in terms of the States of America: Transportation Research Board.
number of inches per mile that a laser, mounted in a Herrmann, T., Adey, B.T., Lüking, J., Schindele, N., and
specialised van, jumps as it is driven across the interstate Hajdin, R., 2008. Total benefit – benefit-cost ratio of
and expressway system. standard interventions for road maintenance. VSS 2004/
2. It is assumed that the relationship between I2 and IRI 714. Switzerland: Swiss Federal Roads Authority (in
can be linearly approximated from the difference in the German).
I2 values that correspond to the IRI values between 5m/ Hufschmid, W., 2002. The condition of cantonal roads gives
km and 6 m/km. cause for serious concern. Strasse und Verkehr, January.
The Swiss Association of Road and Transportation
Experts (in German).
Ihs, A., 2004. The influence of road surface condition on traffic
References safety and ride comfort. In: 6th international conference on
Al-Mansour, A.I. and Sinha, K.C., 1994. Economic analysis managing pavements, 19–24 October, Queensland, Australia.
of effectiveness of pavement preventive maintenance. CD-ROM.
Transportation Research Record, No. 1442, 31–37. Uni- Keeney, R.L. and Raiffa, H., 1993. Decisions with multiple
ted States of America: Transportation Research Board. objectives: Preferences and value tradeoffs. Cambridge,
ASTRA, 2003. NISTRA: Sustainability indicators for road UK: Cambridge University Press.
infrastructure projects. Method report. Switzerland: Swiss Keller, M. and Zbinden, R., 2004. Air destroying material
Federal Roads Authority (ASTRA), Switzerland (in emissions of road traffic 1980–2030. No. 355 in the
German). Environmental Series. Switzerland: Federal Authority
Bennett, C. and Greenwood, I., 2001. Modelling road user for the Environment, Forests and Country.
and environmental effects in HDM-4. Vol. 7 of the Khurshid, M.B., Irfan, M., and Labi, S., 2009. Comparison
Highway Development and Management Manual. of methods for evaluating pavement interventions.
France: PIARC, the World Road Association. Transportation Research Record, No. 2108, 25–36. Uni-
Bester, C., 2003. The effect of road roughness on safety. In: ted States of America: Transportation Research Board.
Proceedings of the Annual Meeting of the Transportation Labi, S. and Sinha, K.C., 2005. Life-cycle evaluation of
Research Board. United States of America: Transporta- flexible pavement preventive maintenance. Journal of
tion Research Board. Transportation Engineering, 131 (10), 744–751.
Bürgi, M., 2002. Basic model of value conservation. Strasse Lamptey, G., Labi, S., and Li, Z., 2008. Decision support for
und Verkehr, October. Switzerland: The Swiss Associa- optimal scheduling of highway pavement preventive
tion of Road and Transportation Experts (in German). maintenance within resurfacing cycle. Decision Support
Cleveland, D., 1987. Effect of resurfacing on highway safety. Systems, 46, 376–387.
State of the art report 6. United States of America: Laube, M., 2001. Traffic behavior and accidents near
Transportation Research Board. highway worksites [online]. In: Proceedings of the 1st
ECOPLAN, 2006. Accident costs in road and rail traffic – Swiss Transport Research Conference (STRC). Available
update for the years 1999 to 2004. Switzerland: Swiss from: http://www.strc.ch/conferences/2001 [Accessed 9
Federal Roads Authority (ASTRA) (in German). February 2009] (in German).
654 B.T. Adey et al.

Lee, C. and Machemehl, R., 2005. Combined traffic signal Saaty, T., 2006. Fundamentals of decision making and priority
control and traffic assignment: Algorithms, implementation theory. Vol. VI of the AHP series. USA: RWS Publications.
and numerical results. USA: Center for Transportation Schandersson, R., 1994. Road pavement condition and traffic
Research, University of Texas. safety: some results and conclusions from the Nordic
LTNZ, 2006. Economic evaluation manual. Vol. 1. New project safety and road surface properties. Nordic Project
Zealand: Land Transport New Zealand. TOVE, VTI Sartryck 231, VTI Sweden.
McLean, J., 1981. Driver speed behavior and rural road Stütze, T., 2004. Economically justified investment value for
alignment design. Traffic Engineering and Control, 22 (4), the maintenance of federal highways. Thesis (PhD).
208–211. Technical University of Berlin, Germany (in German).
Müller, A.G., 2005. Swiss mean values 2004. Operational VSS, 2003a. SN 641 820: Cost-benefit analysis in road traffic.
maintenance costs of highways. Switzerland: Müller AG Switzerland: Swiss Association of Road and Traffic
(in German). Experts (VSS), Technical Committee 2: Planning and
Noyce, D., Bahia, H., Yambo, J., and Kim, G., 2005. Development (in German).
Incorporate road safety into pavement management: VSS, 2003b. SN 640 925b: Pavement maintenance manage-
Maximizing surface friction for road safety. United States ment – Condition assessment and index valuing. Switzer-
of America: Midwest Regional University. land: Swiss Association of Road and Traffic Experts
NPRA, 1997. Veg-grepsprosjekte, samlerapport. Internal (VSS), Technical Committee 7: Maintenance Manage-
report no. 1994. Norway: Norwegian Public Roads ment (in German).
Administration (NPRA) (in Norwegian). VSS, 2003c. SN 641 822: Cost-benefit analysis in road traffic –
O’Brien, L.G., 1989. Evolution and benefits of preventive Time costs in person traffic. Switzerland: Swiss Associa-
maintenance strategies. National Cooperative Highway tion of Road and Traffic Experts (VSS), Technical
Research Program Synthesis of Highway Practice 153. Committee 2: Planning and Development (in German).
United States of America: Transportation Research Board. VSS, 2007. SN 641 823: Cost-benefit analysis in road traffic –
Pichler, W., 1981. Decision model for the choice of the road Time costs in freight traffic. Switzerland: Swiss Associa-
pavement construction. Thesis (Dissertation). TU Graz, tion of Road and Traffic Experts (VSS), Technical
Austria (in German). Committee 2: Planning and development (in German).