ramanathan_2001

7/28/2019 ramanathan_2001

1/9

Journal of Environmental Management (2001) 63, 2735doi:10.1006/jema.2001.0455, available online at http://www.idealibrary.com on

A note on the use of the analytic hierarchyprocess for environmental impact

assessment

R. Ramanathan

Indira Gandhi Institute of Development Research Santosh Nagar, Goregaon (East) Mumbai,

400 065, India

Received 30 March 1999; accepted 27 March 2001

Environmental impact assessment (EIA) is an intrinsically complex multi-dimensional process, involving multiple criteriaand multiple actors. Multi-criteria methods can serve as useful decision aids for carrying out the EIA. This paper proposesthe use of a multi-criteria technique, namely the analytic hierarchy process (AHP), for the purpose. AHP has the flexibilityto combine quantitative and qualitative factors, to handle different groups of actors, to combine the opinions expressed

by many experts, and can help in stakeholder analysis. The main shortcomings of AHP and some modifications to it toovercome the shortcomings are briefly described. Finally, the use of AHP is illustrated for a case study involving socio-economic impact assessment. In this case study, AHP has been used for capturing the perceptions of stakeholders onthe relative severity of different socio-economic impacts, which will help the authorities in prioritizing their environmental

management plan, and can also help in allocating the budget available for mitigating adverse socio-economic impacts.

2001 Academic Press

Keywords: environmental impact assessment, analytic hierarchy process, socio-economic impactassessment.

Introduction

Environmental impact assessment (EIA) is aprocedure for assessing the environmental implica-tion of a decision to enact legislation, to implementpolicies and plans, or to initiate developmentprojects. It has become a widely accepted tool forenvironmental management. It has been definedas a process for identifying the likely conse-quences for the biogeophysical and socio-economicenvironments and for human health and wel-fare of implementing particular activities andfor conveying this information, at a stage whenit can materially affect their decision, to thoseresponsible for sanctioning the proposals (Wath-ern, 1988). The United Nations EnvironmentProgramme has defined it as an examination,

Email: [email protected]

analysis, and assessment of planned activitieswith a view to ensure environmentally sound andsustainable development (UNEP, 1996). Detaileddescription of the general EIA methodology canbe found in UNEP (1996), Sinha (1998) and web-sites such as http://www.ext.nodak.edu/iaia/eialist,www. worldbank.org and www.oneworld.org/iied/resource.

EIA is an intrinsically complex multi-dimensio-nal process. Perhaps because of this complexity,implementation of EIA is not entirely satisfactory

(e.g. Moon, 1998). New innovations and methodolo-gies may be needed to improve the EIA process.In fact, the process of EIA has been evolvingever since it was adopted for analysing the envi-ronmental impacts of developmental projects. Inthis paper, we propose the analytic hierarchyprocess (AHP) to address the need for consider-ing multiple criteria and multiple stakeholdersin EIA.

03014797/01/090027C09 $35.00/0 2001 Academic Press


2/9

28 R. Ramanathan

The analytic hierarchy process

AHP is an intuitive method for formulating andanalyzing decisions. AHP has been applied tonumerous practical problems in the last fewdecades (Shim, 1989). Because of its intuitiveappeal and flexibility, many corporations andgovernments routinely use AHP for making major

policy decisions (Elkarmi and Mustafa, 1993). Abrief discussion of AHP is provided in this section.More detailed description of AHP and applicationissues can be found elsewhere (Saaty, 1980, 2000).

Application of AHP to a decision problem involvesfour steps (see below).

Step 1: structuring of the decision

problem into a hierarchical model

It includes decomposition of the decision probleminto elements according to their common charac-teristics and the formation of a hierarchical modelhaving different levels. Each level in the hierar-chy corresponds to the common characteristic ofthe elements in that level. The topmost level isthe focus of the problem. The intermediate levelscorrespond to criteria and sub-criteria, while thelowest level contains the decision alternatives.Figure 1 gives an illustration for a simple decisionproblem of choosing the best house to buy. The top-most level is the Focus of Goal (Best house to buy).The goal is characterised by several criteria, andthe second level indicates these. The criteria con-

sidered in Figure 1 are Price (P), Location (L) andAge (A). One can think of subdividing the criteria

further if necessary. For example, location maybe sub-divided into transport facilities, entertain-ment facilities, hospital facilities, etc. There canbe more such intermediate levels, but Figure 1illustrates the simplest hierarchy involving goal,criteria and alternatives. The last level representsthe alternatives, which are the different housesfrom among which one or a few have to be chosen.If there are more decision-makers (DMs) (i.e. thepersons from whom the judgements are elicited),then one can introduce a level of DMs just below theGoal. But, for the purpose of Figure 1, we assumeonly one DM.

Step 2: making pair-wise comparisons

and obtaining the judgmental matrix

In this step, the elements of a particular levelare compared pairwise, with respect to a specificelement in the immediate upper level. A judgmen-

tal matrix is formed and used for computing thepriorities of the corresponding elements.

First, criteria are compared pair-wise withrespect to the goal. A judgmental matrix, denotedas A, will be formed using the comparisons. Eachentry aij of the judgmental matrix is formed com-paring the row element Ai with the column ele-ment Aj:

ADaiji,jD1, 2, . . . , the number of criteria.

The comparison of any two criteria Ci and Cj (say

Price and Location) with respect to the goal is madeusing questions of the type: of the two criteria Ci

Best house to buy

Location AgePrice

H2 H3H1

Figure 1. A simple AHP Model.


3/9

AHP for environmental impact assessment 29

and Cj, which is more important1 with respect to a

best house and how much more?.Saaty (2000) suggests the use of a 9-point scale

to transform the verbal judgements into numericalquantities representing the values of aij. The scaleis explained in Table 1.

The entries aij are governed by the followingrules:

aij >0; aijD1/aji; aiiD1 for all i

Because of the above rules, the judgmentalmatrix A is a positive reciprocal pairwise com-parison matrix.

Step 3: local priorities and consistency

of comparisons

Once the judgemental matrix of comparisons ofcriteria with respect to the goal is available, the

local priorities of criteria is obtained and theconsistency of the judgements is determined. Ithas been generally agreed (Saaty, 1980, 2000) thatpriorities of criteria can be estimated by finding theprincipal eigenvector w of the matrix A. That is:

AwDlmaxw

1 Important is not the only word representing the basis ofcomparison. Other words that may be used, depending on thecontext, include preferred, relevant, etc.

When the vector w is normalized, it becomes thevector of priorities of the criteria with respect to thegoal. lmax is the largest eigenvalue of the matrix Aand the corresponding eigenvector w contains onlypositive entries.

The consistency of the judgmental matrix canbe determined by a measure called the consistencyratio (CR), defined as:

CRDCI

RI

where CI is called the consistency index and RI,the Random Index.

CI is defined as:

CIDlmaxn

n1

RI is the consistency index of a randomly generatedreciprocal matrix from the 9-point scale, withreciprocals forced. Saaty (1980, 2000) has providedaverage consistencies (RI values) of randomlygenerated matrices (up to size 1111) for a samplesize of 500. The RI values for matrices of differentsizes are shown in Table 2.

If CR of the matrix is higher, it means that theinput judgements are not consistent, and henceare not reliable. In general, a consistency ratio of010 or less is considered acceptable. If the valueis higher, the judgements may not be reliable andhave to be elicited again.

Table 1. The semantic scale used in AHP

Intensity of Definition Descriptionimportance

1 Equal importance Elements Ai and Aj are equally important3 Weak importance of Ai over Aj Experience and Judgement slightly favour Ai

over Aj5 Essential or strong importance Experience and Judgement strongly favour Ai

over Aj7 Demonstrated importance Ai is very strongly favoured over Aj9 Absolute importance The evidence favouring Ai over Aj is of the

highest possible order of affirmation2, 4, 6, 8 Intermediate When compromise is needed, values between

two adjacent judgements are usedReciprocals of

the above

judgements

If Ai has one of the above judgements assignedto it when compared with Aj, then Aj has the

reciprocal value when compared with Ai

A reasonable assumption

Table 2. The average consistencies of random matrices (The RandomIndexRI-values)

Size 1 2 3 4 5 6 7 8 9 10

RI 000 000 058 090 112 124 132 141 145 149


4/9

30 R. Ramanathan

Using a very similar procedure, the local prior-ities of alternatives with respect to each criterioncan be estimated. For example, when the housesare compared pairwise with respect to Price, thelocal priorities of the houses can be estimated.

Step 4: aggregation of local priorities

Once the local priorities of elements of differentlevels are available as outlined in the previousstep, they are aggregated to obtain final prioritiesof the alternatives. For aggregation, the followingprinciple of hierarchic composition (Saaty, 2000) isused:

Final priority of House H1D

iLocal priority ofH1 with respect

to Ci Local priority ofCi with respect to the goal

1

Note that the above is a simple weightedsummation. The final priorities thus obtainedrepresent the rating of the alternatives in achievingthe focus of the problem. Sample AHP calculationsare illustrated in the Appendix. As per thesecalculations, we find that House H1 is the besthouse to buy for the hypothetical DM.

The usefulness of AHP for EIA

The AHP briefly described above can be potentially

useful for EIA in many ways. AHP is a compen-satory MCDM technique in the sense that it admitstrade-offs among the various elements of the model.Hence it can provide an ideal framework for EIAwhich also involves trade-offs among various envi-ronmental problems and development. AHP helpsto elicit the complex judgements of different expertsin a common platform. It also ensures accuracy inthe sense that it has an inbuilt method to check theinconsistency of judgements. This ensures that the

judgements are provided only with sufficient careand the error due to negligence is thus minimised.

It should be noted that other multi-criteria meth-

ods such as the multi-attribute utility theory (e.g.Keeney and Raiffa, 1993) can also be appliedand have also been applied in similar situations(Keeney, 1979). Both have their advantages anddisadvantages as evident from a series of reg-ular debates in prominent journals (e.g. Saaty,1980; Dyer, 1990a,b; Harker and Vargas, 1990).The advantages of AHP over other multi-criteria

methods, as often cited by its proponents, are itsflexibility, intuitive appeal to the decision-makers(experts and stakeholders here), and its abilityto check the inconsistencies in judgments (Saaty,2000).

Combining qualitative and quantitative

elements

In conventional EIA methods such as checklistor matrix methods, the choice of elements (orsub-elements) is constrained by the availabilityof a suitable measurable indicator. This restric-tion vanishes when AHP is used, as AHP has theability to handle even qualitative attributes (byproviding suitable quantification using a semanticscale) and has the versatility to mix quantitativeand qualitative elements (Wedley, 1990). This isbecause the method can make use of human judge-ments. Sometimes, indicators may be segregated

into measurable and non-measurable, and only thelatter may be employed in the AHP model to gettheir corresponding scores, while for the former theappropriate measures form the respective scores(Ramanathan and Ganesh, 1995).

Aggregation of many expert opinions

As we have seen, EIA requires expert opinions frommultiple actors in terms of multiple criteria. Typi-cally, there will be more than one expert who will beconsulted in each field of impact (such as air, water,

land, noise, aesthetics, socio-economics, etc.), andthere will be several such groups of experts fromdifferent fields. Consulting more experts will avoidbias that may be present when the judgements areconsidered from a single expert. When judgementsfrom many experts are considered, it is necessaryto aggregate them suitably. Several methods areavailable in AHP for performing the aggregationincluding the geometric mean method and arith-metic mean method (Ramanathan and Ganesh,1994; Peniwati, 1996; Saaty, 2000).

Necessity to consider different groups ofexperts

We have seen that EIA requires considerationof expert opinion from many different fields.In such a case, it is important to study theopinions of experts from different fields on acommon platform. Sometimes, weights have to


5/9


be assigned to the opinions of groups of expertsbelonging to different fields. Conventional methodssuch as checklists cannot synthesise such diverseinformation. AHP possesses some models for thepurpose (Ramanathan and Ganesh, 1994), whichcan be advantageously used.

For example, suppose that several groups ofexperts are involved in assessing a particularproject, and that it is desired to assign weightsto the groups. Assignment of such weights isquite difficult, as no group will accept those fixedby an external agency. However, as shown byRamanathan and Ganesh (1994), a participatoryapproach can be adopted. This approach derives theweights of the different groups using intrinsicallyderived ratings of each group, which comparesitself with the other groups. The method has beenapplied to compare different groups of experts whenchoosing the most appropriate energy mix for urbanhouseholds (Ramanathan and Ganesh, 1995).

Participation of stakeholders

The recent disputes on environmentally sensitiveprojects have led to the necessity to considerall the stakeholders (i.e. key actors) of a project(such as the authorities, local and affected people,engineers, and others). Several studies on environ-mentally and socio-economically sensitive projectsconsider such a stakeholder analysis (Grimble andChan, 1995; Grimble and Wellard, 1997; Adger

et al., 1998). The stakeholders and their interests

in the project should first be identified. Proper cor-rective actions, if needed, should be carried out intime for ensuring smooth execution of the project.For example, the opinions of the people affecteddirectly by the project on the impacts they arelikely to face when the project goes on streamshould be seriously considered. Any misconcep-tion by the local people in this regard shouldbe rectified. Timely corrective actions should betaken so that local people feel positively about theproject.

Several methods such as ranking are possible toelicit the subjective opinions of the stakeholders

on the different impacts of the project. However,AHP can be a very valuable tool for the purposeas it can be devised to capture the feelings of thelaymen and convert their feelings to a numericalscale that reflects their thinking. As the thoughts oflaymen may not be very structured, it is necessaryto verify the accuracy of their judgements. This

verification is possible when AHP is used as

the inconsistencies of judgments can be easilyidentified.

Some shortcomings andmodifications of AHP

In spite of its immense popularity, several short-

comings of AHP have been reported in the liter-ature. Several modifications have been suggestedto the original AHP to overcome these shortcom-ings, and it is important that a user of AHP shouldknow them. Hence, we review briefly some of themore obvious shortcomings and modifications inthis section.

Scale

When introducing AHP, Saaty (1980) advocatedthe use of an additive scale ranging from 19(see Table 1). He defended the scale by providingevidence from a variety of sources. However,several alternative scales have been proposedin the literature. One of the most widely citedalternative scales is the geometric scale (Lootsma,1999), which uses the range (e0g to e8g) for thesame semantic descriptions available in Table 1(g is a constant). The argument for using thisgeometric scale is that AHP tries to capture ratioinformation (relative preference of one alternativeover another), and hence one should use a ratiocharacterisation for the purpose.

Methods for the estimation of priorities

Saaty (1980) advocated the use of the eigenvectortechnique for deriving the weights from a givenpairwise comparison matrix. It is possible touse other techniques for the same purpose. Themost often discussed alternative technique isthe Logarithmic Least Squares Technique (LLST)(Crawford and Williams, 1985). LLST tries tochoose those weights that minimise the logarithmicsquared deviations. Crawford and Williams (1985)

have shown that the LLST solution can be easilyobtained by geometric means.

Rank reversal

One of the most controversial issues in the use ofAHP is the rank reversal phenomenon: the ranking


6/9

32 R. Ramanathan

of alternatives determined by the AHP may bealtered by the addition of another alternative forconsideration. Belton and Gear (1983) showed thatthe ranking of a set of three alternatives changeswhen a copy of one of the alternatives is addedto the set. Dyer (1990a) claims that this problemis a symptom of arbitrary rankings provided bythe AHP. Harker and Vargas (1987) claim thatthis problem can be overcome by constructing anetwork (a system in which the elements of a levelare affected by the levels above as well as below it)rather than considering the system as a hierarchy(a system in which the elements of a level areaffected only by the level above it). A more detaileddiscussion of the AHP networks is available inSaaty (2000).

The concept of absolute measurement (Saaty,1987; Chattopadhyay and Ramanathan, 1998), asagainst the relative measurement conventionallyused in AHP models, does not suffer from therank reversal problem of Belton and Gear (1983).In this approach, the AHP is used to assignscores to ratings on the criteria, such as high,average, low, etc. and then alternatives areevaluated by assigning a rating to the performanceof alternatives on each criterion.

MAHP, the multiplicative variant of AHP, doesnot suffer from rank reversals of the type shownby Belton and Gear (1983) (Lootsma, 1999). This isbecause it uses multiplicative operations through-out rather than mixing additive and multiplicativeoperations as done in conventional AHP. MAHPuses the LLST instead of eigenvector technique,and uses a multiplicative aggregation instead of

the simple weighted aggregation of the principleof hierarchical composition. When multiplicativeaggregation is used, (1) is modified as follows:

Final priority of House H1D

Local priority ofH1 with respect

to Ci Local priority ofCi with respect to the goal

More details of the theory and applications ofmultiplicative AHP can be obtained from Lootsma(1993, 1999) and Ramanathan (1999).

Axiomatic framework

One of the earliest criticisms of AHP was of itslack of an axiomatic framework. Saaty (1986) hasprovided the necessary axioms, pertaining to recip-rocal comparisons, homogenity, independence, andexpectations.

Number of comparisons

AHP uses redundant judgements for checkingconsistency, and this can exponentially increasethe number of judgements to be elicited from DMs.For example, to compare eight alternatives on thebasis of one criterion, a total of 28 judgementsis needed. If there are N criteria, then the totalnumber of judgements for comparing alternativeson the basis of all these criteria will be 28N. Thisis often a tiring exercise for the decision-maker.Some methods have been developed to reduce thenumber of judgements needed (e.g. Millet andHarker, 1990).

Having highlighted the benefits of AHP anddiscussed its shortcomings and modifications, wenow discuss a simple application of AHP below.

Application of AHP forsocio-economic impact

assessment: a case-study

Here, a practical application of AHP for socio-economic impact assessment (SEIA) is describedbriefly. SEIA is usually a part of EIA. For exam-ple, in India, the government requires a reha-bilitation plan of a project affecting people asa part of an EIA report before granting envi-ronmental clearance (Ramanathan and Geetha,1998). A more detailed discussion of the case-studyreported here is available in Nag and Ramanathan(1996).

In this case-study, a SEIA for a proposed LPGrecovery plant in an industrially backward area inthe state of Maharashtra has been studied. First ofall, the likely major socio-economic impacts due tothe proposed project were identified by preliminarysurveys. These include housing, transport, watersupply, sanitation and health.

However, the authorities responsible for theproject would like to know not only the potentialsignificant impacts, but also about their relativeimportance. While the experts in the SEIA teamcan reasonably estimate relative importance, itwould be more desirable if the importance as

perceived by the different stakeholders is also pro-vided. This can help the authorities to decidethe suitability of an environment managementplan. However, it is difficult to compare the dif-ferent impacts using any particular measure, asthese are incommensurable. For example, it isnot possible to propose a measure that com-pares the relative severity of impacts on housing


7/9


with impacts on sanitation. Hence, this problemrequires a methodology that captures perceptionsof different people, and for the purpose of takinga decision, the perceptions should be converted toobjective numbers. AHP is readily applicable insuch cases.

The AHP model is shown in Figure 2. Thesecond level lists the stakeholdersthe company,local administration and people. However, localadministration and people in different affectedregions may behave separately, and hence a thirdlevel is introduced to distinguish the geographicallocations. Usar is the village in which the projectwill take place and which has the highest stake.Other villages in the vicinity will also be affected.

Alibaug and Revdanda are the two nearest townswhich will bear some of the socio-economic impactsof the project. These impacts, whose relativeseverities have to be compared, form the lastlevel.

A separate set of surveys was conducted for

the purpose of prioritising the impacts usingthe model. These surveys required the detailedinvolvement of stakeholders. The stakeholders atthe different localities were asked to comparepair-wise the relative severity of impacts and tocomplete a questionnaire. The questionnaire hadto be translated into the local language during theinterview.

From the pair-wise comparisons of the impacts,a judgmental matrix was formed for each stake-holder. This matrix was used for computing the

priorities (which will be proportional to the rel-ative severity) of the impacts, and the usualconsistency check was carried out. The priori-ties expressed by different people in the samestakeholder-group were combined using arithmeticmeans (Ramanathan and Ganesh, 1994).

No attempt was made to assign weights tostakeholders, and the priorities as expressed byeach stakeholder were analysed separately. Itwas found that the priorities expressed by thecompany and the local administration were similarto those expressed by the town people. In a similarway, the perceptions of the local administrationof villages were similar to those of the villagers.People both in towns and in villages have perceivedthe water-supply problem to be the most severeimpact during the construction phase. Town peoplehave considered sanitation to be the next mostsevere impact, followed by housing, transportand finally, health. In the villages, transporthas been expected to suffer the second most

severe impact, followed by housing, sanitation andhealth.

The priorities can also provide an approximateguide for the allocation of total money availablefor mitigating the adverse socio-economic impacts.For example, the AHP exercise indicates that, toget the full co-operation of the project, it may bemore prudent to allocate nearly half the funds (ear-marked for minimising the negative socio-economicimpacts) to improve the water-supply situation ofthe project area.

Relative severity of socio-economic impacts

Localadministration

Company People

User Alibaug RevdandaOther

villagesUser Alibaug Revdanda

Othervillages

HousingWatersupply

Sanitation Health Transport

Figure 2. AHP model for socio-economic impact assessment.


8/9

34 R. Ramanathan

Summary and conclusion

In this paper, several advantages of using the ana-lytic hierarchy process (AHP) as a tool while carry-ing out an environmental impact assessment havebeen highlighted. Some shortcomings and modi-fications have been described briefly. A practicalapplication of AHP for conducting socio-economic

impact assessment has been discussed. In thisapplication, AHP has been used for capturing theperceptions of stakeholders on the relative sever-ity of different socio-economic impacts, which willhelp the authorities in prioritising their environ-mental management plan. Therefore, we concludethat AHP can be a useful tool for systemati-cally analysing the opinions of several groups ofexperts belonging to diverse fields in an environ-mental impact assessment study, and hope that thetechnique will be advantageously employed in envi-ronmental impact assessment studies in future.

References

Adger, W. N., Brown, K., Tompkins, E., Bacon, P.,Shim, D. and Young, K. (1998). Evaluating Options

for Sustainable Coastal Resource Management: AMulti-Stakeholder Approach. Paper presented at thefifth biennial meeting of the International Societyfor Ecological Economics, November 1519, Santiago,Chile.

Belton, V. and Gear, T. (1983). On a shortcoming ofSaatys method of analytic hierarchies. Omega 11,228230.

Chattopadhyay, D. and Ramanathan, R. (1998). A newapproach to evaluate generation capacity bids. IEEETransactions on Power Systems 13, 12321237.

Crawford, G. and Williams, C. (1985). A note on theanalysis of subjective judgment matrices. Journal of

Mathematical Psychology 29, 387405.Dyer, J. S. (1990a). Remarks on the analytic hierarchy

process. Management Science 36, 249258.Dyer, J. S. (1990b). A clarification of Remarks on the

Analytic Hierarchy Process. Management Science 36,274275.

Elkarni, F. and Mustafa, I. (1993). Increasing theutilization of solar energy technologies (SET) inJordan: Analytic Hierarchy Process. Energy Policy 21,978984.

Grimble, R. and Chan, M.-K. (1995). Stakeholder anal-ysis for natural resource management in developingcountries. Natural Resources Forum 19, 113124.

Grimble, R. and Wellard, K. (1997). Stakeholder method-ologies in natural resource management: a reviewof principles, contexts experiences and opportunities.

Agricultural Systems 55, 173193.Harker, P. T. and Vargas, L. G. (1987). The theoryof ratio

scale estimation: Saatys Analytic Hierarchy Process.Management Science 33, 13831403.

Harker, P. T. and Vargas, L. G. (1990). Reply to Remarkson the Analytic Hierarchy Process by J. S. Dyer.

Management Science 36, 269273.Keeney, R. L. (1979). Evaluation of proposed storage

sites. Operations Research 27, 4864.Keeney, R. L. and Raiffa, H. (1993). Decision with

Multiple Objectives: Preferences and Value Trade-offs.Cambridge: Cambridge University Press.

Lootsma, F. A. (1993). Scale sensitivity in the multi-plicative AHP and SMART. Journal of Multi-Criteria

Decision Analysis2

, 87110.Lootsma, F. A. (1999). Multi-criteria Decision Analy-sis via Ratio and Difference Judgement. Dordrecht:Kluwer.

Millet, I. and Harker, P. T. (1990). Globally effectivequestioning in the Analytic Hierarchy Process. Euro-

pean Journal of Operational Research 48, 8897.Moon, B. (1998). Environmental impact assessment

in Queensland, Australia: a governmental massacre!Impact Assessment and Project Appraisal 16, 3347.

Nag, Barnali and Ramanathan, R. (1996). An applicationof AHP for socio-economic impact assessment. In

Proceedings of the Fourth International Symposiumon the Analytic Hierarchy Process, pp. 353362.Simon Fraser University, Burnaby, British Columbia,Canada.

Peniwati, K. (1996). The Analytic Hierarchy Process:The possibility theorem for group decision making. In

Proceedings of the Fourth International Symposiumon the Analytic Hierarchy Process, pp. 202214.Simon Fraser University, Burnaby, British Columbia,Canada.

Ramanathan, R. (1998). A multi-criteria methodo-logy to the global negotiations on climate change.

IEEE Transactions on Systems, Man and Cybernet-icsPart C: Applications and Reviews 28, 541548.

Ramanathan, R. (1999). Selection of appropriate green-house gas mitigation options. Global EnvironmentalChange: Human and Policy Dimensions 9, 203210.

Ramanathan, R. and Ganesh, L. S. (1994). Grouppreference aggregation methods employed in the

AHP: an evaluation and an intrinsic process forderiving members weightages. European Journal ofOperational Research 79, 249265.

Ramanathan, R. and Ganesh, L. S. (1995). Energyresource allocation incorporating quantitative andqualitative criteria: an integrated model using goalprogramming and AHP. Socio Economic Planning

Sciences 29, 197218.Ramanathan, R. and Geetha, S. (1998). Socio-economic

impact assessment of industrial projects in India.Impact Assessment and Project Appraisal 16, 2732.

Saaty, T. L. (1980). The Analytic Hierarchy Process:Planning, Priority Setting and Resource Allocation.New York: McGraw-Hill.

Saaty, T. L. (1986). Axiomatic foundations of the analytic

hierarchy process. Management Science32

, 841855.Saaty, T. L. (1987). Rank generation, preservation andreversal in the analytic hierarchy process. Decision

Sciences 18, 157177.Saaty, T. L. (1990). An exposition of the AHP in reply to

the paper: remarks on the analytic hierarchy process.Management Science 36, 259268.

Saaty, T. L. (2000). Fundamentals of Decision Makingand Priority Theory with The Analytic Hierarchy

Process. Pittsburg: RWS Publications.


9/9


Shim, J. P. (1989). Bibliographical research on theanalytic hierarchy process (AHP). Socio-Economic

Planning Sciences 23, 161167.Sinha, S. (1998). Environmental impact assessment: an

effective management tool. TERI Information Monitoron Environmental Science 3, 17.

UNEP (1996).Environmental Impact Assessment: Issues,Trends and Practice. United Nations EnvironmentProgramme, Nairobi.

Wathern, P. (1988). An introductory guide to EIA.

In Environmental Impact Assessment: Theory andPractice (P. Wathern, ed.), pp. 330. London: UnwinHyman.

Wedley, W. C. (1990). Combining qualitative and quan-titative factors: an analytic hierarchy approach. Socio-

Economic Planning Sciences 24, 5764.

Appendix

Illustration of calculations for the AHP modelshown in Figure 1 using hypothetical data

Table A1. Comparison of criteria with respect to theoverall objective

Price Location Age Local priorities

Price 1 3 5 0637Location 1/3 1 3 0258

Age 1/5 1/3 1 0105

lmaxD3039; CID0019; CRD0033

Table A2. Comparison of the three houses with respectto Price

H1 H2 H3 Local priorities

H1 1 4 6 0691H2 1/4 1 3 0218H3 1/6 1/3 1 0091


Table A3. Comparison of the three houses with respectto Location


H1 1 3 5 0637H2 1/3 1 3 0258H3 1/5 1/3 1 0105


Table A4. Comparison of the three houses with respectto Age


H1 1 5 4 0674H2 1/5 1 1/3 0101H3 1/4 3 1 0226


Table A5. Final priorities of the three houses

Final priorities

H1 0675H2 0216H3 0109

ramanathan_2001

Documents