Available online at (Elixir International Journal) Civil Engineering. Elixir Civil Engg. 80 (2015)

Transcription

1 30 Alireza Arabameri / Elixir Civil Engg. 80 (0) Available online at (Elixir International Journal) Civil Engineering Elixir Civil Engg. 80 (0) Appliation of One of Multi Attribute Deision Making Models for Site Seletion of Damp for the Purpose of Sustainable Development Case Study: (BAKHTIARI Basine) Alireza Arabameri Department of Civil Engineering, College of Engineering, Buinzahra Branh, Islami Azad University, Buinzahra, Iran. ARTICLE INFO Artile history: Reeived: 7 Otober 0; Reeived in revised form: 8 February 0; Aepted: 6 Marh 0; Keywords Damp, ELECTRE Method, Site Seletion, Sustainable Development. ABSTRACT In reent deades, duo to inrease in population, demand for reliable water supplies has inreased. The situation is ritial in plaes where groundwater is the only aessible water resoure and the disharge rate of water is more than the rate of reharge. In suh areas, artifiial reharge of groundwater is an important management strategy. In establishing an artifiial reharge sheme, site Seletion is the prime prerequisite and its suess depends on the olletion and analysis of a great deal of geographi data. The use of ELECTRE in site seletion projets allows the deision-mars to inorporate unquantifiable information into deision model. In new researhes of site seletion multiple riteria deision- mars methods have been onsidered. In this paper, site seletion of artifiial groundwater reharges using an integrated method of ELECTRE has been arried out by raster layers in GIS. The result and findings of Researh show that in ELECTRE method, zones (3,) dominated () times and defeated () time, so it is loated in the first rank with () points and is the most suitable zone for establish damp. In ontrast, zone () defeated (6) time and dominated no time, therefore it is loated in the last rank with (-6) points and is not the most suitable zone for artifiial reharge. And, zones (, 6, 7,,) dominated (,,, ) times and defeated (,,,,) and loated in other ranks with (, -, -, -) points respetively. Also, zones (7, 6,, ) should be omitted beause their defeated times are more than dominated times. 0 Elixir All rights reserved. Introdution Sustainability and a resilient ity is one of the important hallenges of humanity in the twenty first entury. In other words, today the main world opportunities and hallenges are embedded in ities and the rapid growth of urbanization aompanied by industrial ativities led to ineffiieny of urban infrastrutures and an intensive inrease of environmental destrution[0]. In the urrent situation, assessing environmental sustainability is one of the most important tools in the proess of planning for sustainable development. This assessment is a type of eologial assessment whih would be arried out in different levels in sequential way to present a framework for analyzing and assessing the impats of plans, strategies and polies on environment in a omprehensive method by offering some reommendations to mitigate environmental pressures. So, providing an appropriate milieu for assessment and measuring environmental sustainability is inevitable in the proess of urban development and planning[]. Water resoures management is a set of various management ativities aimed at the optimum utilization of water resoures and redution of eonomial, soial and environmental damages and losses. Deision making issue in water resoures management is very omplex and ompliated beause of several deision indiators and riteria [9]. Ahieving a determine purpose, there are a lot of solutions with different priorities for various issues suh as environmental, soial, organizational and politial problems. These neessities leads to use of multiple riteria deision making aimed at seletion of best solution among different solutions. In previous deades, deision making in water management problems and seletion of better option among suggested options to solve a watershed problems was only done based on eonomial riteria - profit in relation to ost- and on hanging soial and environmental riteria in to the eonomial riterion. However, today using Multi riteria deision making, it is not neessary to use finanial equivalent of soial and environmental riteria to selet the best option. In fat, various qualitative and quantitative riteria an be used to prioritize and selet the best options for water resoures management. Eletrial Method is a type of available methods in Compensatory Models. In this method whole options evaluate by non-rand omparisons. All stages of this method are established based on oordinated and unoordinated sets and thus this method is known as Coordination Analysis. Banayoun established the Eletrial Method and Delft, Nijkamp, Roy and their olleagues developed it. In Eletrial method, the onept of domination uses impliitly. In this method, options are ompared in pairs, then dominant and weak (dominant and defeated) options determined and weak or defeated options omitted [8]. There are several studies on ground water and their artifiial reharge all over the world. Noori et al (00) tried to find the appropriate areas for artifiial reharge of ground water by reharge pools and GIS tehnique in watershed Gavbandi and introdued alluvial fans and pediplain as the best area for artifiial reharge[7]. Mousavi et al (00) found the potential appropriate areas for artifiial Tele: addresses: alireza.ameri9@yahoo.om 0 Elixir All rights reserved

2 303 Alireza Arabameri / Elixir Civil Engg. 80 (0) reharge of ground water in the viinity of Kamestan antiline by integration of remote sensing and GIS tehniques and introdued bron formations, alluviums and river anals as the best position for artifiial reharge[]. Mianabadi and Afshar (008) investigated and rand the projet of water supply in Zahedan using three methods: Indued Ordered Weighted Averaging (IOWA), Linear Assignment and TOPSIS methods, and then they ompared the findings of these methods with the results of adaptable planning method []. Limon and Martinez (006) used Multi Attribute Utility theory for optimum alloation of agriulture water in north of Spain [6]. Ahmadi et al (00) used multiple riteria deision making to rank different projets of refining agriulture water to reuse them []. Also, Anand Raj and Kumar (996) rand management options of river basin by ELECTRE method []. There are many examples of appliations of Multi Criteria Deision Making in literature (For instane: The evaluation of servie quality []; Inter ompany omparison []; The appliations in aggregate prodution planning [6], Faility loation seletion [7] and large sale nonlinear programming [8]. The purpose of this study is zoning the best area for artifiial reharge of underground basins in BAKHTIARI watershed using effetive fators in reharging underground water table by ELECTER method. In another way, this study aimed at the seletion of most appropriate area to establish damp for the purpose of sustainable development of water resoures using Multi Criteria Deision Making methods (ELECTRE) and lassify the best areas in onsidered watershed. Methods and materials Mathematial situation of studied area Being situated in the enter part of Lorestan provine, Charmahal and Bakhtiari is bounded by 3º, 09 latitude to 3º, 8 north latitude and 9º, 8 to º and longitude. It has aess to Lorestan provine in north, to Khozestan provines in south, to Isfahan provine in east and to Charmahal and Bakhtiari provine in west. Globally, Bakhtiari is loated at 80 meter height above sea level. Figure. Mathematial situation of studied area Researh Methodology Firstly, studied area was investigated by the satellite images of Google Earth and its limitations were determined. Then digital elevation model of area was separated from its digital elevation model in Iran in the environment of soft ware GlOBAL MAPER and the output was reeived. Required data layers for zoning in the environment of software Ar GIS 9.3 was prepared as following: First, digital elevation model lassified in to 7 elevation lasses based on natural breaks in the heights of the area. Mentioned lasses represent the studied zones in the area and subsequent alulations were done in eah of these lasses. Slope layer prepared base on digital elevation model on the area by surfae analyses tool in 3D analyses. There were different proesses to prepare drainage density layer and habitual density suh as digitizing main and minor waterways layers on the topographial map:0000 of the area, digitizing main and minor fault on geologial map :00000 of area and density tool in Spatial Analyses. Iso-Preipitation layer prepared by interpolating method li ringing tehnique and linear relationship between rain-height using Interpolate tools in 3D analyses. Seond, the investigated riteria for eah height zones were alulated (Tables ) and their layers prepared separately. After ahieving a few numbers in eah layer, the numbers were analyzed by ELECTRE method. Then onsidered watershed was rand to selet the best area for establishing damp. Applying ELECTRE Tehnique for site seletion of damp. Establishing Deision Making Matrix: Aording to the riteria and numbers of options and evaluation of whole options for the different riteria, Deision Making Matrix develops as follow; x x n X. x m xmn In whih the Funtion of Xij (i =,,., M) is in relation to the riteria I j (j =,,3,, n).. Sale down the Deision Making Matrix In this stage, all riteria with different dimensions is hanged into the dimensionless riteria and matrix R defined as follows. There are several methods to sale down, but generally the following equation used in eletrial method [3]. r r n xij R. r ij m r m rmn xij i () 3. Determining Weighted Matrix of riteria w W 0 w 0 w n As you an see, Weighted Matrix (W) is diagonal matrix in whih the elements on main diameter are not zero and amount of these elements equal to importane oeffiient of the related vetor.. Determining Weighted Normalized Deision Matrix Weighted Normalized Deision Matrix is obtained by multiplying Sale down Deision Making Matrix into the Weighted Matrix of riteria. v v V R W v m vmn. Establishing agree and disagree riteria set The riteria set J = (, m) divides into two subsets; agree and disagree for eah pair of options e, k (k, e =,,., M, k # e). Agree Set (S Ke ) is a set of riteria in whih option K is preferred to option e. and its omplementary set is the opposite set (I Ke ) in mathematial language; S j v v n ()

3 30 I j v v 6. Establishing Agree Matrix: To establish agree matrix, its elements, agree indiators, should be alulated. Agree indiator is sum of weight of riteria in agree set. Thus, indiator C is between option k and option e equals to [8]: W js j W For total normalized weights j j W j j s : W j j Alireza Arabameri / Elixir Civil Engg. 80 (0) equals so: : )( Agreement represents the superiority of options k on option e whih its amount hanges in the range of zero to one (0-). After alulating agree indiator for all options, matrix whih is a m * m matrix is defined as follows. Generally, this matrix is not symmetrial. C m m( m ) m m 7. Determining Opposite Matrix Disagreement indiator (opposite) is desribed as follows [9]. max v v d ji max v v jj (6) The amount of disagreement indiator hanges from zero to one. After alulating disagree indiator for all options, matrix whih is a m * m matrix is defined as follows. Generally, this matrix is not symmetrial. d d m d dm D dm dm( m ) It notied that the data inluding in agreement matrix, are different from data in opposite matrix and in fat these data are ompleted eah other. The differene between the weight is developed through agreement matrixes, while the differene between determined values is obtained through opposition matrix. 8. Establishing agree dominant matrix: In the sixth step, it indiated how to alulate agreement indiator C. Now there is a determined amount for agreement indiator in this step whih is alled agreement threshold C is larger, option k is preferred on option e, otherwise it is not. Agreed threshold is alulated by the following equation. [8] m m k e m( m ) k e ek : (7) Agree Dominated Matrix (F) is developed based on the amount of agreement threshold and its elements determined in the equation bellow [] (3) )(. If f 0 (8) 9. Establishing Opposed Dominane Matrix : Opposed Dominane Matrix (G) is established the same as Agree Dominated Matrix. First, deision mars should express opposite threshold whih is for example the mean of opposite indiators (disagreement) [8]. m m d d k e m( m ) k e ek (9) Similar to seventh step, it is better that the amount of opposite indiator (d ) beome less, beause opposite amount (disagreement) expresses superiorities dimension of option k on option is aeptable. In ontrast, if (d ) were larger than, opposite amount would be very great and it would not be ignored. Thus, Opposed Dominane Matrix is defined as follows [8] 0 d d g d d (0) Eah element in the matrix (G) shows the dominant relationship between options. 0. Establishing Final Dominant Matrix: Final Dominant Matrix (H) is developed after multiplying eah element in Agree Dominated Matrix (F) into elements in Opposed Dominane Matrix (G) [8]. h f. g (). Removing less satisfation options and seleting the best option: Final Dominant Matrix (H) indiates detail preferenes of options. For example, when amount of h equals, it means that option k is preferred on option e in both agree and disagree situation (it means its preferene is larger than the agree threshold and its opposite or weakness is less than disagree threshold), but option k may be dominated by other options yet. The options should be rand in a way that the more dominated options are seleted than the more defeated one. Determining the importane oeffiient of options than the other, riteria are ompared in pair by time suggested method. Table. Weighting the fators based on preferene in paired omparison [3] Numerial values Preferenes (judging verbal) 9 Extremely preferred 7 Very strongly preferred Strongly preferred 3 Moderately referred Equally preferred 8 6 Intervals between strong preferenes After the formation of paired omparison matrix, relative weights of riteria an be alulated. There are different methods to alulate the relative weight based on paired omparison matrix. The most important ones are the "least squares method, least squares logarithmi method, speial vetor method and approximate method. The speial vetor method is the most aurate one. In this method, Wi is determine in the equation: A W=λmaxW () In this equation, λ and W are speial amount and speial vetor of paired matrix respetively. If dimensions of matrix were

4 30 Alireza Arabameri / Elixir Civil Engg. 80 (0) Table. Deision Matrix (X) Table 3. Sale down Deision Matrix (R) Table. Paired Comparison Matrix of different riteria (S) Criteria Materials Preipitation Stream density Slope fault density Elevation area Wij Materials / 3868 Preipitation / 39 Stream density / 8 Slope / 08 fault density / 0603 Elevation / 033 area / 0 Inonsisteny rate: 0/0 (due to being less than 0/ ompatibility matrix indies are aeptable) Table. Weighted Normalized Deision Matrix (V) Table 6. Agreement Matrix (C)

5 306 Alireza Arabameri / Elixir Civil Engg. 80 (0) Table 7. Opposite Matrix (D) Table 8. Agree Dominated Matrix (F) Table 9. Opposite Dominated Matrix (G) Table 0. Final Dominated Matrix (H) Table. Number of dominant and reessive of eah seleted areas Differene Number being defeated Rule number Regions

6 307 Alireza Arabameri / Elixir Civil Engg. 80 (0) larger, alulation would be too time onsuming. So, to alulate λ, the amount of Dtrmynal λia-matrix will be equaled to zero. Considering the greatest value of λ in equation (3), the amount of wi is alulated. [] A λ max.i = 0 (3) Disuss In reent years, water exploitation has beome greater for many reasons suh as population growth, industrial development, urbanization growth and onsequently inreased demand for food produts. Hene the rate of exploitation and onsumption ground water beome greater than reharge of them, in other words input of ground water system is less than its output and system with negative balane sheet has positive feedbak and it is ollapsing. Thus it is very signifiant to determine and assign the suitable position for this ase. Water resoures management is a set of various management ativities aimed at the optimum utilization of water resoures and redution of eonomial, soial and environmental damages and losses. Deision making issue in water resoures management is very omplex and ompliated beause of several deision indiators and riteria. Ahieving a determine purpose, there are a lot of solutions with different priorities for various issues suh as environmental, soial, organizational and politial problems. These neessities leads to use of multiple riteria deision making aimed at seletion of best solution among different solutions. Sustainable development is the management and onservation of basi natural resoures and diretion of tehnial and organizational hanges to ahieve and prepare requirements for generations at the present and in future. Suh a development in agriulture setion leads to the onservation of water, soil and plants and it is nondestrutive environmentally, proper tehnially, frugal eonomially and aeptable soially. Similar to under development ountries, our ountry needs to ompress and develop agriulture in order to arry out enormous requirements of under growth population. Thus, it is neessary to aquire the exat and up to date information about the ondition of water resoures and predition of their situation in future in order to ahieve optimum management for water resoures. One of the management methods for water resoures is Multi Criteria Deision Making. In reent deades, several researhers attempt to use Multi Criteria Deision Making in omplex and ompliated deisions. These deision methods divide into two parts; Multi Objetive Deision Making, Multi Attribute Deision Making. Multi Criteria Models use to selet the best options. Evaluative Models for MADM lassify into two models; Compensatory Model, Non- Compensatory Model. Non-ompensatory model inludes methods whih don`t need to ahieve data from DM and lead to objetive answer. Exhanging between inditors is permitted in Compensatory model. It means that for example, a weakness in a indiator may be ompensated by option of other indiator. Eletrial Method is a type of available methods in Compensatory Models. In this method whole options evaluate by non-rand omparisons. All stages of this method are established based on oordinated and unoordinated sets and thus this method is known as Coordination Analysis. Banayoun established the Eletrial Method and Delft, Nijkamp, Roy and their olleagues developed it. In Eletrial method, the onept of domination uses impliitly. In this method, options are ompared in pairs, then dominant and weak (dominant and defeated) options determined and weak or defeated options omitted []. Linear Assignment is one of the Multi Criteria Deision Making ombines qualitative and quantitative indiators, weights riteria based on their importane and helps deision mars to selet the best options at the same time. In this method, supposed options are rand based on their points in eah available indiator and then the final rank of the options determined by the Linear Compensatory Proess. The results of ELECTRE method to find the most suitable area for artifiial reharge of groundwater aquifers of BAKHTIARI watershed showed in tables () to (). Therefore, a matrix is formed with rank (9) for data matrix, with 7 alternatives (height zones) and 7 related indiators (Materials, Preipitation, stream density, fault density, slope, Elevation) (Table ). Conlusion The inreasing trend in (urban) water demand due to population growth plaes a growing stress on available water resoures and alls for an effiient and aeptable long-term management of the resoures. Hene, appliation of multiattribute deision-making systems is essential for evaluating urban water supply shemes. A number of multi-attribute deision-making methods have been developed. This paper aims to survey the appliation of suh systems to urban water supply problems and the effets of eah multi-attribute deision-making method seleted on the final ranking of alternatives. Three methods of Indued Ordered Weighted Averaging (IOWA), ELECTRE and TOPSIS have been onsidered for a real urban water management ase study in the ity of Bakhtiari in Iran. The results revealed that the multi-attribute deision-making method seleted had a onsiderable effet on the final ranking of a finite set of alternatives suh that different MADM tehniques yielded different results for the same problem. It is, therefore, neessary to selet the method aording to the speifi harateristis of the problem at hand, type of data available, and the assessments made. The ultimate alternative must be, thus, seleted one evaluations have been made of the results obtained from applying different deision-making methods to the problem Referenes [] Ahmad,S.A,Tewfik,S.R,Talaa,H.A (00). Development and Verifiation if a Deision Support System For The Seletion if optimum water reuse Sheme, Desalination,V,PP [] Anand,Raj,P.A., Kumar, D.N. (996). Ranking of niver basin alternative using ELECTRE, Hydrologial Siene, V,PP36-33 [3] Krishnamurthy, J., N. Kumar, V. Jayaraman& M. Manivel, (996). An Approah to Demarate Ground Water Potential Zones Thorough Remote Sensing and a Geographial Information System, INT. J. Remote Sensing, 7 (0): [] Mianabadi, H, Afshar, A, (008). Multi attribute Deision Making to rank urban water supply Sheme, water and watershed journal, v9, n66, pp 3 [] Mousavi, S.F; Chitsazan, M; Mirzaei, Y; Shaan, M; B Mohammadi, H.R, (00), Integrating remote sensing and GIS to find potential suitable areas for ground water: Kamestan Antiline area, Conferene artiles and Geomanti exhibition [6] Limon,G.A,Martinez, Y, (006), Multi0riteria modeling of irrigation water Mard at basin level : aspsnish Case Study, Eropiangeornal of operational of Researh, V73,PP [7] Noori, B.; Ghayoumian, J.; Saravi Mohseni, M; Darvish- Sefat, A.A; Feiz-Nia, Sadat,00, determining suitable areas for artifiial reharge of groundwater in reharge pools using GIS, Journal of Natural Foundations of Iran, Volume 7, Number 3, pp [8] Roy,B, (99),The outranking Approah and the foundation of ELECTRE Methods, Theory and Deision,3, pp 9-73.

7 308 Alireza Arabameri / Elixir Civil Engg. 80 (0) [9] TiLLe,M, Dumont, A.G., (003). Methods of Multi riteria Deision Analysis Within the Road Projet li an Element of the Sustainability, 3 rd Swiss Transport Researh Conferene,Marh 9-. [0] Tavakoli, A.Ra, and A.R Ali Ahmadi, (006), Seletion and prioritization model of transferring methods in Azad Shahr, Rey, Tehran, Siene and Tehnology University Press [] Saaty, T.L., (99), Highlights and ritial points in the theory and appliation of the analytial hierarhy proess, European Journal of operational researh, Vol. 7, Pp. 6-7 [] Ghodsi Poor, H., (009), Analysis Hierarhal Proess, Tehran, Amir Kabir University Press, Fifth Edition [3] Vami, (99), Projet opportunity Study on Integrated use of the Razgah Nepheline ores, Iran by metallurgial proessing into Alumina Cement, sodium Carbonate and potash, final report, Volume, general explanatory note,99 [] Tsuar, S.H., Chang, T.Y., Yen, C.H., (003). The evaluation of airline servie quality by fuzzy MCDM, Tourism Management, 3, pp. 07. [] Deng, H., Yeh, C.H., Willis, R.J., (000). Inter-ompany omparison using modified TOPSIS with objetive weights, Computers &Operations Researh, 7, pp [6] Wang, R.C., Liang, T.F., (00). Appliation of fuzzy multi-objetive linear programming to aggregate prodution planning,computers & Industrial Engineering, 6, pp. 7. [7] Chu, T.C., (00), Faility loation seletion using fuzzy TOPSIS undergroup deisions, International Journal of Unertainty, Fuzzinessand Knowledge-Based Systems, 00, 0(6), pp [8] Abo-Sina, M.A., Amer, A.H. (00), Extensions of TOPSIS for multiobjetive large-sale nonlinear programming problems, Applied Mathematis and Computation, 00, Artile in press.