工程招標外文翻譯--多準則支持招標(節(jié)選)_第1頁
已閱讀1頁,還剩13頁未讀, 繼續(xù)免費閱讀

下載本文檔

版權說明:本文檔由用戶提供并上傳,收益歸屬內容提供方,若內容存在侵權,請進行舉報或認領

文檔簡介

1、<p>  中文4350字,2900英文單詞,16500英文字符</p><p>  文獻出處:Seydel J, Olson D L. Multicriteria support for construction bidding[J]. Mathematical & Computer Modelling, 2001, 34(5):677-701.</p><p>  M

2、ulticriteria Support for Construction Bidding</p><p>  J.SEYDEL, D.L.OLSON</p><p>  Abstract-While profit maximization is one important objective in this decision domain, other objectives are im

3、portant as well. This paper discusses multiple criteria and their respective objectives in construction bidding, and presents a bidding framework which recommends a pairwise comparison procedure to generate criterion wei

4、ghts and a linear transformation procedure to calculate relative scores for bidding alternatives. This hybrid multicriteria method is illustrated and evaluated using set </p><p>  Keywords-Bidding Constructi

5、on, Multicriteria analysis, Decision theory.</p><p>  1. INTRODUCTION</p><p>  Competitive bidding presents numerous tradeoffs to those who submit bids for work on construction projects. If a bi

6、d is relatively high, then the probability the bid will be accepted is relatively low, thus resulting in: low expected revenues, low equipment and personnel utilization, and the opportunity for a competitor to build his/

7、her standing in the industry. However, such a bid, should it be accepted, will also result in higher profits, less chance of a loss resulting from unforeseen costs, </p><p>  This research should be of inter

8、est to the many decision makers (DMs) who face these tradeoffs regularly. Nearly all public (and a substantial amount of residential) construction projects involve competitive bidding, and construction is a major industr

9、y throughout the world. In the United States, for example, construction activity accounted for nearly five percent of the country’s gross national product in the 1980s according to reports of the Bureau of Census [1] and

10、 the Economic Report of the </p><p>  1.1 Multicriteria Nature of Bidding</p><p>  Multiple criteria involved in bidding have been discussed for decades. In the first known work formalizing bidd

11、ing optimization, Friedman addressed the existence of multiple bidding criteria by listing objectives of profit maximization, maximizing return on investment, minimization of loss expectation, minimizing competitor profi

12、ts, and maximizing operational continuity [4]. Boughton addressed these multiple objectives as well [6]. Not unexpectedly, he found, in a survey of 126 construction firm</p><p>  Actual applications of multi

13、criteria analysis to competitive bidding are limited, however. Engelbrecht-Wiggans developed a descriptive model analyzing the simultaneous maximization of profit and minimization of two forms of regret [9]. The first ge

14、neral prescriptive applications of multicriteria methodology in competitive bidding are found in [l&12]. Ahmad proposed a twostage approach, based on multiattribute utility theory (MAUT), for the decision of whether

15、or not to bid on a project, and then</p><p>  1.2 Scope of Consideration</p><p>  Besides the studies indicated above, there have been numerous other studies addressing competitive bidding in a

16、variety of applications, as discussed in an extensive survey by Engelbrecht-Wiggans [18]. To aid in the analysis of the many existing bidding applications, King and Mercer developed a classification scheme to summarize p

17、roblems addressed by this body of literature [19].According to this scheme, bidding situations are classified according to four factors: whether bidding is open (as in </p><p>  2. GENERAL BIDDING MODEL</

18、p><p>  Let the decision maker of concern be referred to herein as the subject bidder, and let bids being submitted by the subject bidder’s competitors be referred to as competing bids. Then, for the subject bi

19、dder, the bidding decision can be modelled as an unconstrained optimization problem. This problem is stochastic in that the outcome depends upon the values of two random variables-the lowest competing bid amount, and the

20、 actual cost or value of the object for which the bid is being submitted .If a</p><p>  subject to resource limitations, where </p><p>  M (the decision variable) is the markup ratio, of bid

21、to bidder’s estimated cost Ce</p><p>  C (a random variable) is the ratio of actual to estimated cost,</p><p>  ML (a random variable) is the ratio of the lowest competing bid to the bidder

22、’s cost estimate,</p><p>  Y(C, ML, M) is the outcome on the given decision criterion, ?A(C, M) is the outcome given the bid is accepted (bidder is successful), N is the outcome given the bid is not accepte

23、d (bidder is unsuccessful), and Pr(Win〡M) is the bid acceptance probability. </p><p>  Both the decision space (values of M) and the joint probability sample space (combinations of ML and C) are continuous.

24、</p><p>  In practice, however, these spaces contain finite sets of positive numbers somewhere near one.</p><p>  The distributions of ML and C are generally not known, but must be estimated fro

25、m empirical data. Several terms are expressed as ratios rather than as absolute measures, because the only link among historical data used to generate estimates for the distribution of ML is the cost estimate C,. Further

26、more, dividing through by C, results in relationships that are independent of the cost estimates. While ML and C, are not necessarily independent, empirical studies by Hansmann and Rivett [22] and Kuh</p><p>

27、;  The probability of winning, Pr(Win 1 M), can be modelled as the probability of beating the lowest competing bid, following Hansmann and Rivett [22], Carr and Sandahl [24], Gunter and Swanson [25], and Mercer and Russe

28、ll [26]. </p><p>  Required data are the bidder’s estimated costs for past projects, as well as the lowest competing bids for those projects. </p><p>  Prom this information, probability distrib

29、utions can be estimated. </p><p>  Since Pr(Win 1 M) = Pr(ML > M), the acceptance probability for a bid amount M is equal to 1 - FL(M), and the nonacceptance probability is FL(M), where FL(M) is the cumul

30、ative distribution function of ML evaluated at M. </p><p>  In most of the bidding literature, the N term reflecting the outcome--essentially, the cost-of an unsuccessful bid does not appear. This is likely

31、because the firm ordinarily receives no revenue for losing a contract, and prior costs for preparing the bid are typically sunk. The nonacceptance term is incorporated here to facilitate the ability to accommodate circum

32、stances in which it might be desirable to model nonacceptance outcomes explicitly. Such circumstances would occur when it is possible</p><p>  2.1Multicriteria Model</p><p>  When the above mode

33、l is modified to incorporate multiple criteria, the problem can be expressed in terms of the multiple objectives</p><p>  subject to resource and policy limitations, where m is the number of decision criteri

34、a. This can alternatively be expressed in terms of a, the decision maker’s multiattribute value function</p><p>  for i = 1 to m and subject to resource and policy limitations. The Ai and Ni are defined in t

35、erms of the multiple objective formulation. If the criteria exhibit additive independence, then E[R(C,ML,M)] becomes a weighted sum of the single criterion utility functions evaluated at the Yi(C,M~,M) as in [9]. I n suc

36、h cases, multicriteria methods such as multiattribute value (MAV) analysis, and possibly AHP, can be appropriate for solving the problem, although neither of those is without its own prob</p><p>  3. SUMMARY

37、</p><p>  While numerous works have addressed bidding optimization over the past 40 years, industry practitioners generally continue to ignore those works and, instead, use rules of thumb and other arbitrary

38、 approaches for determining prices [6,19]. This likely results, at least in part, from the fact that existing optimization methods tend to produce unsatisfactory results, as they are incapable of incorporating tradeoffs

39、among multiple criteria. Furthermore, in order to be successfully adopted by indust</p><p>  Based upon actual data, a probability model has been developed to take into consideration major factors which affe

40、ct bidder behavior. Then, based upon that model, a multicriteria optimization approach was presented and evaluated for effectiveness. That approach employs an elicitation procedure (preferably based on pairwise compariso

41、ns) to generate estimates of the weights for the DM’s multicriteria objective function. In the evaluation of the proposed bidding optimization approach, a wide range</p><p>  For a good portion of the prefe

42、rence structures considered, the multicriteria approach, when compared to the profit-based approach, led to substantial or nearly substantial improvement in DM value outcomes. On the other hand, for a few preference stru

43、ctures, there was little improvement when the additional criteria were considered. Essentially, in situations where there was relatively little concern about regret (“money left on the table”) and related criteria, the p

44、roposed multicriteria optimiz</p><p>  An extension of this research would consider the automation of the process, and hence, address a third component for the bidding system: an information processing capab

45、ility which would make it possible to acquire, maintain, and manipulate the data required by bidding optimization. One such system is already operational and is described by Hegazy and Moselhi [15] and Moselhi et al. [17

46、]. In addition, somewhat different architectures have been proposed by Ahmad and Minkarah [14] and Seydel [35], </p><p>  Another extension of this research must address how to modify the proposed system, if

47、 possible, to deal with situations in which bid takers (i.e., potential customers) are awarding contracts according to criteria beyond price (i.e., lowest bid). Increasingly, firms are seeking to implement Deming’s fourt

48、h point: “end the practice of awarding business on price tag alone” [37].Customers are looking at ways to choose trade partners on the basis of product quality, provider reputation, delivery spee</p><p><

49、b>  多準則支持招標</b></p><p>  J.SEYDEL,D.L.OLSON</p><p><b>  摘 要</b></p><p>  利潤最大化是一個領域的重要目標,其它目標也是同樣重要的。本文討論了多個工程招標的標準和各自的目標,并提出了一個框架,建議采用一個兩兩對比的招標以便于產生評標標準和線性轉換的

50、過程,用于計算投標方案的相對分值。這種混合型的方法是用于說明和評價過去采用的一套工程招標集。擬定的招投標制是當工程量被發(fā)現成為大幅度提高預期利潤時非常重要的解決方案。函數中確認利潤減少后,為了使得提高多屬性功能可能實現,進行評估預期利潤損失,(確認函數中的利潤減少后評估預期利潤損失)。</p><p>  關鍵詞:招標;施工;多目標分析;決策理論</p><p><b>  1引

51、言</b></p><p>  競標為那些提交建設項目工程的投標單位提供了眾多權衡。如果投標價相對較高,那么其被接受的概率就會相對較低,從而會導致:低的預期收入,低的設備和人員利用率,以及為競爭對手在本行業(yè)建立他/她的地位的機會。然而,這樣的投標價假如被接受的話,將會帶來更高的利潤,不可預見的機會成本所造成的損失也會減少,并能夠為客戶提供更高水平的質量。因此,多重準則都受到所確定的投標金額的影響,而且

52、存在著嚴重的權衡取舍需要考慮。本論文中闡述的研究目的是要努力開發(fā)出一種可行的、合用的、以施工為導向的系統(tǒng)來提供一體化的背景因素和決策標準,它由全面可靠的數據來支持。最終目標是為建筑行業(yè)的決策者在定價公司服務時,提供有效的多重準則支持。朝著這一目標,本文描述、演示并評估了一個多因素量化程序和多標準優(yōu)化投標。</p><p>  許多經常面對這些權衡的決策者(DMs)應該對本研究有興趣。幾乎所有的公共場所(和大量的住

53、宅)建設項目涉及競爭性招標,建筑業(yè)是一個遍布世界各地的重要產業(yè)。例如在美國,根據聯邦調查局對人口普查的報告[1]和美國總統(tǒng)經濟報告[2],20世紀80年代美國的建筑活動幾乎占了國民生產總值的5%。正因為如此,大量的文獻致力于競爭性投標各方面的研究,既有從描述性的角度也有從說明性的角度[3]。不難理解,競爭性投標的有效性隊許多公司的未來是至關重要的,假如出價過高,就會降低中標的可能性,然而出價太低,則會導致財政失敗。自從弗里德曼[4]于1

54、957年首次開發(fā)了定量招投標優(yōu)化模型后,大量的模型已經被提出來支持招投標。然而,正如羅特科普夫和哈爾斯塔[3]已指出的,從業(yè)者已較少使用這些模型了。他們認為,理論和實踐之間的差距通常是,考慮到語境的豐富模型的需要與走向現實主義之間的差距。同樣地,Rothkopf和EngelbrechtWiggans[5]已經指出,有缺陷的模型會有錯誤的結果。那么理所當然地,只為追求利潤最大化的模型將很難被實際投標人所接受,他們是典型的多重標準關注者。&

55、lt;/p><p>  1.1招標的多準則性質</p><p>  招標中涉及的多重準則已經過了數十年的討論。在第一個已知的正式優(yōu)化招標工作中,弗里德曼指出,利潤最大化的上市目標,投資回報最大化,預期損失最小化,競爭對手利潤最小化,連續(xù)性經營最大化,使多重招標準則得以存在[4]。鮑也指出了這些多重目標[6]。果不其然,他在對126家建筑公司的調查中發(fā)現,利潤最大化雖然決不是唯一重要的目標,但卻

56、是招標中最常使用的目標。卡爾指出了投資和生產標準回報率,雖然利潤最大化還是納入他模型的唯一目標[7]。在之后的研究中,卡爾將機會成本納入了以利潤為基礎的招標方法[8]。最后,盡管在文獻中沒有明確提及,中標是許多承包商(包括作者和作者所見過的公司)默認的目標。這個目標也許能或也許不能代表其他目標的集合,如資源利用率和現金流量的維持,但無論如何不能也不應該在實際招投標實踐中忽視它。</p><p>  然而,多準則分

57、析在競爭性投標中的實際應用是有限的。Engelbrecht-Wiggans開發(fā)了一個描述性的模型,分析了利潤最大化和兩種損失形式損失最小化的同步發(fā)生[9]。第一個多準則方法在競標中的一般規(guī)定性應用被發(fā)現?;诙鄬傩孕в美碚?MAUT),艾哈曼德就是否決定投標一個項目,并且之后應該使用什么樣的標注提出了二級辦法。不幸的是,效用函數的啟發(fā)作用,尤其是多屬性效用函數,對DMs來說是復雜且耗時的。由賽德爾和奧爾森提出的確定最佳標注的方法是基于薩

58、蒂[13]引進的層次分析法(AHP)的一部分。計算得到簡化,DM的負荷也減少了,盡管比起艾哈邁德的多屬性效用方法,該方法依靠的是更嚴格的假設。隨后,艾哈邁德,Minkarah[14]和Hegazy等[15-17]將他們這些行之有效的多準則招標方法的各個方面實施在計算機軟件中。</p><p><b>  1.2審議范圍</b></p><p>  除了上述研究表明,已

59、經有其他不少研究應用于競爭性招標,如En--gelbrecht-Wiggans的廣泛調查論[18]。為了幫助現有的招投標分析,國王和得瑟制定了通過總結和分類文學本身解決辦法的計劃[19]。根據這個計劃,招標按照四個因素進行分類:公開招標(例如房地產拍賣)或議標(密閉招標)、招標選擇(投標價高,低價中標或者其他)、非價格競爭的存在(所有投標人滿足同樣的條件,或提出替代產品),并且該項目的確定值稱為買入后價值。研究的重點是建筑招投標活動,是

60、在其中使用密封招標,低價中標價格已經選定,所有投標人達到同樣的標準,項目成本是確定的。(要注意的是,盡管越來越多除價格以外的因素被合同授予,合同授予的最低投標人仍沒被淘汰。例如投標人的資格預審和嚴格規(guī)范,尤其是在公共建設工程,歷來擔任代表非價格屬性。本質上,這些標準代表限制,不僅僅是招標過程中的8s目標。雖然有被納入作為實現這些諸多好處的目標的其他標準,這樣做是對未來研究的保留。)本研究的目的的提供一個由Seydel和Olson發(fā)展的擴

61、展方法,通過集合到一個通用的多框架,并通過招標和評標的框架說明實際施工招標使用的數據。注意的是,招標</p><p><b>  2一般招標模型</b></p><p>  在此處讓決策者關注的是統(tǒng)稱為專業(yè)招標者,并讓專業(yè)招標者的競爭對手提交投標價被稱為競爭性招標。然后,作為專業(yè)招標者,招標決策可以作為一個無約束的優(yōu)化問題來建模。這個問題是隨機的,最低投標量的競爭的結

62、果取決于兩個隨機變量的值,和正在提交投標的實際成本或價值規(guī)定的對象。如果用單一的標準考慮,其目標是優(yōu)化招投標過程的結果?;诟ダ锏侣墓ぷ鱗4],優(yōu)化的典型方法是尋求預期的優(yōu)化結果(例如利潤)。一般低價中標的情況表明,目的是確定投標比率M,以</p><p><b>  受資源限制,在這里</b></p><p>  M是(決策變量)招標對投標人成本估計Ce標注比率;

63、</p><p>  C是(隨機變量)實際估計成本比率;</p><p>  ML是(隨機變量)競爭性最低的招標對投標人成本估計比率;</p><p>  Y(C, ML, M) 是在給定決策標準上得出的結果;</p><p>  A(C, M)是給出可接受的投標人的結果(中標);</p><p>  N是不可接受投標

64、人的結果(未中標);</p><p>  Pr(Win〡M)是中標的概率.</p><p>  決策空間(M的價值)和可能聯合的試樣空間(ML和C的結合)都是連續(xù)的。然而在實踐上,這些空間的地方附近包含了有限正數集合。</p><p>  ML和C的分布一般是未知的,必須從經驗數據里估計。幾個作為比率的條件不是絕對的,因為用于生成ML分布的歷史數據的唯一聯系時成本估

65、計Ce。此外,通過被Ce除,導致了獨立的成本估計的聯系。Hansmann and Rivett [22] 和 Kuhlmann 和Johnson [23]實證研究ML和C沒有必要的聯系。因此,相關數據應該在特設及基礎上分析,在下一節(jié)會作簡要討論,以確定此類的獨立程度。</p><p>  中標的概率Pr(Win〡M),可以作為擊敗低價中標的模型,Hansmann 和 Ri--vett [22],Carr 和 Sa

66、ndahl [24],Gunter 和 Swanson [25], and Mercer 和Rus--sell [26]。</p><p>  所需的數據是招標人對先前項目的成本估計,以及這些項目的最低競爭投標。</p><p>  從這些信息,可以估計概率分布。</p><p>  因Pr(Win〡M)= Pr(ML > M),投標價M接受的概率等同于1-F

67、L(M),不被接受的概率是FL(M),在這里FL(M)是ML評估M的累積分布函數。</p><p>  在大多數投標文獻中,沒有出現反映不成功投標的結果(本質上是成本)的N項。這很可能是因為該公司通常不會因失去合同而獲得收入,而且準備投標的前期成本通常會下降。此處引入不接受項是為了便于適應可能需要明確地對不接受結果進行建模的情況。當有可能估計不成功的投標結果時,就會出現這種情況,包括但不限于次優(yōu)(即計劃B )備選

68、方案;與替換在緩慢時期離開公司的關鍵人員有關的費用;搬遷費用;以及解散公司的費用。為了更容易地遵循要演示的過程,除了在下一節(jié)中包括在通用多標準模型中之外,在此不明確考慮這些成本。也就是說,為了演示的目的,假設N具有零值。盡管如此,一些不接受費用,例如與關鍵人員損失有關的費用,雖然不是通過同時考慮多項標準來處理,如下文所述。當然,未來的研究是有必要的,在識別和建模不接受的結果,因為它們確實存在,投標人知道它們,但招標優(yōu)化文獻一般忽略它們。

69、</p><p>  2.1多準則模型當上述模型被修改為包含多個標準時,問題可以用多個目標來表示</p><p>  受資源和策略限制,其中m是決策標準的數量。這也可以用決策者的多屬性值函數來表示</p><p>  i = 1 to m,并受到資源和政策的限制。A的i次和N的i次從多準則目標的制定中被定義。如果附加的標準表現獨立性,那么E[R(C,ML,M)]

70、在評定Yi(C,M~,M)時成為一個單一的標準實用程序函數的加權綜合,比如在公式[9]中。在這種情況下,多準則方法比如多準則價值(MAV)分析和可能的層次分析法,可以成為合適的解決問題的方法,盡管這些方法并不是沒有自身的問題。</p><p><b>  3小結</b></p><p>  在過去的40年,雖然許多的著作已經處理了投標優(yōu)化的問題,但業(yè)界普遍繼續(xù)無視這些

71、著作,相反,他們使用經驗規(guī)則或其他任意的方法確定價格[6,19]。這至少部分可能是由于現有的優(yōu)化方法無法在多個標準中納入權衡,往往產生無法令人滿意的結果。此外,為了能成功地被業(yè)界采用,投標優(yōu)化,不管多準則與否,都必須作為系統(tǒng)的一部分。這樣的一個系統(tǒng)必然是由至少兩部分組成的:在評估各種投標方案中所需的概率信息生成的分析組件和在給出所涉及到的權衡中確定哪一個是最好的優(yōu)化組件。盡管分析組件已在本文中簡要討論過,在其他工作中也詳細探討過[5,2

72、4,32],優(yōu)化組件是本文的重點。它的成功取決于分析組件始于保持良好的記錄,之后實施健全統(tǒng)計程序的可靠性。</p><p>  該系統(tǒng)的觀點指導著本文的研究,討論已在這里列出?;趯嶋H數據,一種概率模型已經被開發(fā)出來,用來考慮影響投標人行為的主要因素。然后,基于該模型,多準則的優(yōu)化方法被提出并對其有效性進行評估。這種方法采用非啟發(fā)式程序(最好是在兩兩比較的基礎上)為DM的多準則目標函數生成權重估計。在提出的投標優(yōu)

73、化方法評估中,考慮到了廣泛的偏好結構(正如MAV函數所表明的)。對所考慮的很大一部分的偏好結構來說,當與以利潤為基礎的方法比較時,多準則方法在DM價值結果中會產生實質性的或近乎實質性的改善。另一方面,對一些偏好結構來說,當考慮到附加標準時,很少有改善。從本質上說,關注優(yōu)化及相關標準的情況不多(“錢放在桌上”),多準則優(yōu)化方法證明是有效的。</p><p>  研究延伸的一個方面是過程的自動化,因此,對招投標系統(tǒng)的

74、三部分確定:可能獲取的信息處理能力、維護、通過招標優(yōu)化操作必要的數據。一個這樣的系統(tǒng)已經開始運行,由Hegazy、Moselhi[15]和Moselhi[17]等進行描述。此外,被Ahmad 和Minkarah [14]、Seydel [35]、Ward和Chapman[36]建議用有些不同的體系結構用于提供招標信息系統(tǒng)的一般準則。其他擴展的研究包括一個更詳細的切實考慮承包商的標準,以及研究通過成對比較標準權重的確定和通過其他程序的決定

溫馨提示

  • 1. 本站所有資源如無特殊說明,都需要本地電腦安裝OFFICE2007和PDF閱讀器。圖紙軟件為CAD,CAXA,PROE,UG,SolidWorks等.壓縮文件請下載最新的WinRAR軟件解壓。
  • 2. 本站的文檔不包含任何第三方提供的附件圖紙等,如果需要附件,請聯系上傳者。文件的所有權益歸上傳用戶所有。
  • 3. 本站RAR壓縮包中若帶圖紙,網頁內容里面會有圖紙預覽,若沒有圖紙預覽就沒有圖紙。
  • 4. 未經權益所有人同意不得將文件中的內容挪作商業(yè)或盈利用途。
  • 5. 眾賞文庫僅提供信息存儲空間,僅對用戶上傳內容的表現方式做保護處理,對用戶上傳分享的文檔內容本身不做任何修改或編輯,并不能對任何下載內容負責。
  • 6. 下載文件中如有侵權或不適當內容,請與我們聯系,我們立即糾正。
  • 7. 本站不保證下載資源的準確性、安全性和完整性, 同時也不承擔用戶因使用這些下載資源對自己和他人造成任何形式的傷害或損失。

評論

0/150

提交評論