Key Risk Factors Assessment for Metropolitan

Transcription

1 Key Risk Factors Assessmet for Metroolita Udergroud Project Shih-Tog Lu 1 Cheg-Wei Li 2 Hsi-Lug Liu 3 1 Deartmet of Risk Maagemet, Kaia Uiversity, Taiwa 2 Deartmet of Logistics ad Shiig Maagemet, Kaia Uiversity, Taiwa 3 Detartmet of Civil Egieerig, Mighsi Uiversity of Sciece & Techology, Taiwa Abstract Udergroud costructio roject i metroolis is more dyamic ad risky. A key risk factors aalysis will give roject cotractor a more ratioal basis o which to make decisio. This study alies the fuzzy referece relatios ad icomlete liguistic referece relatios to deal with the degree of imact ad rak for the mai risk factors of a metroolita udergroud roject. Keywords: Metroolita udergroud roject, Risk factors aalysis, Fuzzy referece relatios, Icomlete liguistic referece relatios 1. Itroductio Because of the ecoomic growth, i order to satisfy the eeds of the eole, the urba area eeds to ivest for the corresodig ublic costructios. But due to the limitatio of the lad availability i the metroolita/urba area, may ublic costructios are develoed toward udergroud, such as the MRT, subways or sewers ad so o. Also, the dyamic, risk, variatio, challege i executig udergroud ublic costructios are much higher tha the commo o-urba/rural costructios. Esecially for the crowded oulatio, cogested traffic ad dese ielies, costructio udergroud ublic works will be qualified as high-risk costructios. If the cotractors ca t rovide a suitable assessmet ad rocessig of the risk ad ucertaity factors of the metroolita udergroud costructios, it may result i time delay, over-budget ad oor erformace [1][2]. Next, it would cause loss to the cotractor ad ower ublic-related sector). That is, the stakeholders of the roject will have loss i fiace ad reutatio due to the deficiet risk assessmet ad maagemet strategies. Uder this circumstace, it is a must to have a overall assessmet of the risk ad ucertaity factors before executig udergroud metroolita ublic works. The determiatio of sigificat risk factors ad rovidig ecessary maagemet strategies to overcome these key risk factors is a big icetive to each of the stakeholders of the rojects. I other words, how to let the key risk factors of the udergroud ublic works udergo a rimary assessmet ad la-out a aroriate maagemet strategy is oe imortat work for the executio of this roject. This research utilizes the literature review to adot a well established structure of the risk factors of the udergroud roject executio. After which, we itroduce the cocet of cosistet fuzzy referece relatios ad eve use umerical examles to lot out the rocess method of assessig the 1

2 roject s key risk factors. Lastly, we rovide a discussio of the roblems observed ad a coclusio. 2. Assessmet of key risk factors This sectio s goal is to describe a key risk factors assessmet aroach of the ublic udergroud costructio that icludes two arts: 1) itroducig risk factors of the executio of the rojects o udergroud ublic works ad 2) determiig the degree of imact of the key risk factors Itroducig the risk factors structure of a udergroud roject This study maily establishes a risk assessmet method o the executio of ublic udergroud works. Also, with the results from the literatures, Ghosh ad Jitaaakaot [3] used factor aalysis aroach o risk factors of Thailad s udergroud works ad created a corresodig overall aalysis. Thus this study adots the established risk factors of the research as basic assessmet for the key risk factors of metroolita ublic udergroud works. These risk factors iclude 9 comoets: fiacial ad ecoomic risk, cotractual ad legal risk, subcotractor-related risk, oeratioal risk, desig risk, safety ad social risk, force majeure risk, hysical risk, ad delay risk Determiig the degree of imact of risk factors The degree of imact here is based o the cocet of Zhi [4] who shows the degree of seriousess whe ivolutary thigs hae ad the scale of imact they cause o the roject. Because each risk factor of the roject has differet imortace ad imlicatio, therefore, we caot assume the same results from each risk factor's degree of imact. May researchers aly the aalytical hierarchy rocess AHP) which was develoed by Saaty [5] [6] as their aalysis tool to solve multi-criteria decisio roblems. But, whe the criteria ad factors are much more, the airwise comariso used will create comlicatio ad it is hard to have cosistecy for the researchers. Herrera-Viedma et al [7] roosed the fuzzy referece relatios to solve the icosistecy of the decisio matrices of airwise comarisos based o additive trasitivity. Xu [8] roosed a oeratioal laws of liguistic evaluatio scale to deal with icomlete liguistic referece relatios also by usig the additive trasitivity roerty. Because this study has to evaluate a lot of risk factors, therefore, we use the two aroaches to calculate the weighs of the risk factors as the degree of imact o the roject case of the risk factors. We rovide a basic itroductio o the defiitio ad stes of the two roosed methods which is show below: Cosistet fuzzy referece relatios A multilicative referece relatio A o a set of alteratives/criteria X is rereseted by a matrix A " X! X, where A = a ), a is the referece itesity ratio of alterative/criteria x i to alterative/criteria x j. Saaty [5][6] suggested measurig a to be scaled from 1 to 9. The assessed values ad corresodig meaig are show i table 1. 2

3 Table 1. Liguistic scale of the assessmet of the degree of imact of risk factors i+ 1) i+ 2) j"! i i+ 1) = j " i + 1) / 2,! i < j 1) j + ji 1) Degree of Imortace Values Equal Imortace 1 Weak Imortace 3 Essetial Imortace 5 Very Strog Imortace 7 Absolute Imortace 9 Itermediate Values I this case, the referece relatio A is tyically assumed to be multilicative recirocal, a! a = 1 " i, j {1,!, }.The fuzzy ji referece relatio P geerated by X is a fuzzy set of X! X, that is characterized by a membershi fuctio µ : X " X! [0,1]. The referece relatio may be coveietly rereseted by the! matrix, P = P ), where = µ x, x ) " i, j!{1,!, }, i j is the degree of referece ratio of alteratives/criteria x i ad x j. Herei, = 1/ 2 meas x i ~ x j ; > 1/ 2 meas x i! x j etc., ad + ji = 1 for i, j!{1,!, }. Suose there have a set of alteratives/criteria X = { x 1,! x }, ad associated with the -1 referece itesity ratio, a,!, } of { a a 12 23! 1 alterative X for a![1/ 9, 9]. The corresodig recirocal fuzzy referece relatios P ),![0, 1] ca be obtaied by = = g a ) = 1/ 2! 1 + log a ). The use 9 the followig formula 1) to obtai the other referece relatios values of ot belogig to{ 12, 23,!,! 1}. However, all the ecessary elemets i the decisio matrix will ot all lie withi [0,1] but will lie withi [! a, 1+ a], where a = mi{! i, j}. Therefore the decisio matrix with etries that ca be obtaied by trasformig the obtaied values usig a trasformatio fuctio that maitais recirocity ad additive cosistecy. The trasformatio fuctio is: x + a f :[! a, 1 + a] " [0,1], f x) = 2) 1 + 2a The obtaied assessmet decisio matrix shows the cosistet recirocal relatio P! =! ). Fially, it ca use the formula 3) to obtai the imortace weight) of each factor: A = 1/ "! ), W = A / " A 3) i j= 1 i i i= 1 i Cosistet icomlete liguistic referece relatios Let S = {! = ",!,! } s t t be a fiite ad totally ordered discrete term set, whose cardiality value is odd oe, such as 7 ad 9.[?] Each term, s i reresets a ossible value for a liguistic variable ad has characteristics of s > s if ad! " oly if! > ". S ca be defied as: 3

4 " s = absolute less imortat,! 4 % s = strog less imortat,! 3 s = essetial less imortat,! 2 s = weak less imortat,! 1 $ s equalimortat, & 0 S = = s = weak imortat, 1 s = essetialimortat, 2 s = strog imortat, 3 s = absolute imortat ' 4 The decisio maker comares each air of alteratives/criteria X = { x 1,! x } by usig the discrete term set S, ad costructs a liguistic referece relatio A = a ), where a! idicates the referece degree or itesity for the alterative/criteria x i over x j, ad a! S, a " a = s, a = s for all ji 0 ii 0, i,j. If the decisio maker is uable to rovide referece values for all airs of alteratives/criteria, the A is called a icomlete liguistic referece relatio. If A was costructed a accetable icomlete liguistic referece relatio with oly -1 judgmets a 12, a 23,!, a! 1, the, we ca use the kow elemets i A ad formula 4) to determie all the other ukow elemets i A, ad thus get a cosistet comlete liguistic referece relatio A " = " ). a! a = a! a, for all i, j, k 4) ik kj Fially, it ca fuse all the referece degrees a! j = 1,2,!, ) i the ith lie of the A! by usig the liguistic averagig oerator as formula 5), ad the get the averaged oe a i of the ith alterative/criterio over all the other alteratives/criteria ai = ai! 1 " ai! 2 "!" ai!, for all i 5) Lastly, it ca base o this assessmet results to uderstad the rak of the imortace o the alteratives/criteria. 3. Numerical examles This study alies the two aroaches to assess the imortace of the 9 key risk factors o udergroud roject of Ghosh ad Jitaaakaot [3]: F 1 )fiacial ad ecoomic risk, F 2 )cotractual ad legal risk, F 3 )subcotractor-related risk, F 4 )oeratioal risk, F 5 )desig risk, F 6 )safety ad social risk, F 7 )force majeure risk, F 8 )hysical risk, F 9 )delay risk. As the first aroach, it has to iquire the roject maagers who have te years exeriece o the local costructio to rakly comare the degree of imact imortace) of the 9 risk factors i Taiei s udergroud roject. Namely F 1 comare with F 2, F 2 comare with F 3, F 3 comare with F 4 to make iferece. Lastly, we comare F 8 ad F 9. All these eed to comare 8 referece relatios ad the corresodig comlemets that are show below: A = F F 2 1/7 1 1/5 F F 4 1/4 1 1/6 F F 6 1/7 1 1/7 F F 8 1/5 1 3 F 9 1/3 1 Fig. 1: Fuzzy referece relatios of risk factors ordered airwise comarisos 4

5 For the secod aroach, the results of Fig. 1 will be chaged to the outcomes of Fig. 2. For examle, a = a = 7! a = a = s, F a = 1 5 " a = s 23 23!, a = 4! a = s, P= F a = 1 6 " a = s 45 45!, F a = 1 7 " a = s 67 67!, a = 5 " a = s !, F a = 3! a = s F F 1 S 0 S 3 F 2 S -3 S 0 S -2 F 3 S 2 S 0 S 1.5 F 4 S - S 0 S - A= F 5 S 2.5 S 0 S 3 F 6 S -3 S 0 S -3 F 7 S 3 S 0 S 2 F 8 S -2 S 0 S 1 F 9 S -1 S 0 Fig. 2: Icomlete liguistic referece relatios value of risk factors ordered airwise comarisos F 1 F 2 F 3 F Fig. 3: Recirocal fuzzy referece relatios of risk factors From the matrix above, we ca clearly observe that there are some values that lies outside the rage [0,1]. Therefore, we eed to use formula 2) to rocess trasformatio i order to guaratee the recirocity ad additive cosistecy of the whole matrix. The trasformed cosistet recirocal fuzzy referece relatio matrix is as follows: F F F From the ordered airwise comariso of Fig. 1, we comute their referece relatios value a by the fuctio F = g a ) = 1/ 2! 1 + log 9 a ) ad the P = F formula 1). The a ca be trasformed F to, where lies withi [0, 1] as F show i the results below: F F Fig. 4: Fial decisio matrix of the first aroach Lastly we aly formula 3) to obtai the corresodig degree of imact weights) ad rak the imortace of the risk factors see Table 2.) 5

6 Table 2. The imortace of the risk factors for the first aroach Risk Factor Degree of Imact Rak fiacial ad ecoomic risk F cotractual ) ad legal risk F subcotractor-related ) risk F oeratioal ) risk F 4 ) desig risk F 5 ) safety ad social risk F 6 ) force majeure risk F 7 ) hysical risk F 8 ) delay risk F 9 ) For the secod aroach, utilize the kow elemets i Fig. 2 ad formula 4) to determie all the ukow elemets i A. We get the corresodig cosistet comlete liguistic referece relatios as followig: A' = F 1 S 0 S 3 S 1 S 2.5 S 0 S 3 S 0 S 2 S 3 F 2 S -3 S 0 S -2 S - S -3 S 0 S -3 S -1 S 0 F 3 S -1 S 2 S 0 S 1.5 S -1 S 2 S -1 S 1 S 2 F 4 S - S 0.5 S - S 0 S - S 0.5 S - S - S - F 5 S 0 S 3 S 1 S 2.5 S 0 S 3 S 0 S 2 S 3 F 6 S 3 S 0 S -2 S - S -3 S 0 S -3 S -1 S 0 F 7 S 0 S 3 S 1 S 2.5 S 0 S 3 S 0 S 2 S 3 F 8 S -2 S 1 S -1 S 0.5 S -2 S 1 S -2 S 0 S 1 F 9 S -3 S 0 S -2 S - S -3 S 0 S -3 S -1 S 0 Fig. 5: Fial decisio matrix of the secod aroach Utilize formula 5) to fuse all the referece degrees a! j = 1, 2,!, 9) i the ith lie of the A!, ad the get the averaged oe a of the ith risk factors over all the other risk factors i a1 = s1.61 a2 = s! 1.39 a3 = s0.61 a4 = s! 0.89 a5 = s1.61 a6 = s! 1.39 a7 = s1.61 a8 = s! 0.39 a9 = s! 1.39 Rak all the risk factors i accordace with the values of a i : F! F! F " F " F " F " F! F! F From the simulated values above, we ca ote that after obtaiig the assessmet of exerieced exerts, we ca fid that there are somewhat differet betwee the two aroaches. For the first aroach, the desig risk ad force majeure risk are the highest degree of imact o Taiei s udergroud roject. This result is the same as the secod aroach, but cosiders the fiacial ad ecoomic risk also to be the most imortat. Furthermore, both aroaches cosider the lowest degree of imact is the delay risk. Therefore, if we execute a udergroud roject i Taiei, we ca more or less give attetios o the most sigificat factors before the costructio, so as to come u with good risk cotrol strategies. Because of time ad age limit, this study is t able to assess the occurrece robability of each risk factor. Therefore, we ca iclude this art i future research ad after that itegrate with these results to calculate the risk level or risk degree of each risk factor. 4. Coclusio I the traditioal AHP comarisos, if the there are items have to assess, we eed to comare they i!1) / 2 times. After Herrera-Viedma et al. [7] ad Xu [8] roosed the method of fuzzy referece relatios ad icomlete liguistic referece relatios, the 6

7 umber of airwise comarisos ca be reduced by 1 times. Not oly does it simlify the desigig ad aswerig of the questioaires, but also it ca maitai referece cosistece. There is o eed to sed extra time to solve or ivestigate the questio of cosistecy. Esecially, whe assessig a lot of risk factors or criteria, a better result will be the obtaied. Therefore, this study utilizes this two assessmet methods so that those who are ufamiliar with the comlicated rocess of AHP will be give a big ease. issues o cosistecy of fuzzy referece relatios, Euroea Joural of Oeratioal Research, 154, [8] Xu, Z., 2006, Icomlete liguistic referece relatios ad their fusio, Iformatio fusio, 7, 3, Refereces [1] Kagari, R. ad Riggs, L.S., 1989, Costructio Risk Assessmet by Liguistics, IEEE Trasactios o Egieerig Maagemet, 36, 2, [2] Thomso, P. A. ad Perry, J. G., 1993, Egieerig Costructio Risk: A Guide to Project Risk Aalysis ad Risk Maagemet, Thomas Telford, Lodo. [3] Ghosh, S., Jitaaakaot, J., 2004, Idetifyig ad assessig the critical risk factors i a udergroud rail roject i Thailad: a factor aalysis aroach, Iteratioal Joural of Project Maagemet, 22, 8, [4] Zhi, H., 1995, Risk maagemet for overseas costructio roject, Iteratioal Joural of Project Maagemet, 13, 4, [5] Saaty, T. L., 1980, The Aalytic Hierarchy Process, McGraw-Hill, New York. [6] Saaty, T. L., 1997, A Scalig Method for Priorities i Hierarchical Structures, Joural of Mathematical Psychology, 15, 3, [7] Herrera-Viedma, E., Herrera, E., Chiclaa, F.,Luque, M., 2004, Some 7