2022年工程管理专业外文翻译

1、外文资料翻译资料来源：文章名:Predicting Effectiveness of Construction Project Management: Decision-Support Tool for Competitive Bidding书刊名：An International Journal作 者：Rasa Apanaviciene, Arvydas Juodis出版社：国际杂志，章 节：Vol.6, No.3 / September - December 页 码：P347P360文 章 译 名： 建设工程项目管理的预测功能： 用于决策支持工具竞争性招标 姓 名： 学 号： 指引教师(职

2、称)： 专 业： 班 级： 所 在 学 院： 外文原文Predicting Effectiveness of Construction Project Management: Decision-Support Tool for Competitive BiddingIntroductionConstruction projects are delivered under conditions of risk in the competitive market environment. The origin of risk is the uncertainty inherent to any p

3、roject, and every risk is associated with a cause, a consequence and the probability or likelihood of the event occurring. There are external risks (economic, political, financial and environmental) and internal risks based on project management issues, i.e. projects managers and his team competency

4、, experience, strategic and tactic decisions made during construction project delivery. The opportunity to improve organizational performance through more effective project management could provide substantial savings for construction management company.Project management effectiveness depends on ce

5、rtain factors of project management system. The literature review revealed a substantial volume of work on measuring or identifying the factors or conditions contributing to the effectiveness of construction projects. There are three main trends of previous research on construction project success f

6、actors:key factors identification for construction project success Jaselskis et. A1.(1991); Sanvido et. A1. (1992); Chua et. A1. (1997);identification of key success factors for a particular group of construction projects, e.g.BOT, design-build, public-private partnerships Tiong (1996);Molenaar et.

7、A1. (); Chan et. AI. (), Zhang (), Shen et. A1.();analysis of a particular factor impact on construction project success Cheng et. A1. (); Bower et. A1. (); Ford ().Some writers were attempting to develop predictive models while others focused on generating a list of practices. Predictive models dev

8、eloped to identify the key factors and to measure their impact on overall project success were using regression and correlation techniques, factor analysis, Monte-Carlo simulation, experts and multicriteria decision-making support methods. Essentially in these approaches the functional relationships

9、 between the input factors and project outcome is assumed and tested against the data. The relationships are modified and retested until the models that best fit the data are found.When developing construction project management effectiveness model (CPMEM) referred to here, the writers attempted to

10、cull the best aspects of artificial neural networks (ANN) methodology. The neural network approach does not require an a priori assumption of the functional relationship. Artificial neural networks are very useful because of their functional mapping properties and the ability to learn from examples.

11、 Networks have been compared with many other functional approximation systems and found to be competitive in terms of accuracy Haykin 1999. This and the ability to learn from examples allow modelling the complex construction project management system where behavioural rules are not known in detail a

12、nd are difficult to analyze correctly.Methodology of Artificial Neural NetworksThe foundation of the artificial neural networks (ANN) paradigm was laid in the 1950s, and ANN has gained significant attention in the past decade because of the development of more powerful hardware and neural algorithms

13、 Haykin (1999). Artificial neural networks have been studied and explored by many researchers where they have been used, applied, and manipulated in almost every field. For example, they have been used in system modelling and identification, control, pattern recognition, speech pronunciation, system

14、 classifications, medical diagnosis as well as in prediction, computer vision, and hardware implementations. As in civil engineering and management applications, neural networks have been employed in different studies. Some of these studies cover the mathematical modelling of non-linear structural m

15、aterials, damage detection, non-destructive analysis, earthquake classification, dynamical system modelling, system identifications, and structural control of linear and non-linear systems, construction productivity modelling, construction technology evaluation, cost estimation, organisational effec

16、tiveness modelling and others Adeli et. A1. (1998), Sinha et. A1. (). A neural network can be defined as a model of reasoning based on human brain Wasserman (1993). Learning is a fundamental and essential characteristic of biological neural networks. The ease with which they can learn led to attempt

17、s to emulate a biological network in a computer.2.1 Model of Artificial Neural NetworkAn artificial neural network consists of a number of very simple and highly interconnected processors, also called neurons, which are analogous to the biological neurons in the brain. The neurons are connected by w

18、eighted links passing signals from one neuron to another. Each neuron receives a number of input signals through its connections; however, it never produces more than a single output signal. The output signal is transmitted through the neurons outgoing connection (corresponding to the biological axo

19、n). The outgoing connection, in turn, splits into a number of branches that transmit the same signal (the signal is not divided among these branches in any way). The outgoing branches terminate at the incoming connections of other neurons in the network. Figure 1 represents connections of a typical

20、ANN.As shown in Figure 1, a typical ANN is made up of a hierarchy of layers, and the neurons in the networks are arranged along these layers. Each layer in a multilayer neural network has its own specific function. The input layer accepts input signals from the outside world and distributes them to

21、all neurons in the hidden layer. These neurons detect the features; the weights of the neurons represent the features hidden in the input patterns. These features are then used by the output layer for determining the output pattern. The output layer accepts output signals from the hidden layer and e

22、stablishes the output pattern of the entire network. The neurons are connected by links, and each link has a numerical weight associated with it. Weights are the basic means of long-term memory in ANN. Weights express the strength (importance) of each neuron input. A neural network learns through re

23、peated adjustment of these weights.The network in Figure 1 is fully connected and has a feedforward structure, meaning there are no connection loops that would allow outputs to feed back to their inputs, although a recurrent neural network has feedback loops from its outputs to its inputs. The indic

24、es i, j and k in Figure 1 refer to neurons in input, hidden and output layers, respectively. Input signals, x1, x2 . xi, xn, are propagated from left to right, and error signals, c1, c2 . ci, from right to left. The symbol wij denotes the weight for the connection between neuron i in the input layer

25、 and neuron j in the hidden layer, and the symbol wjk the weight between neuron j in the hidden layer and neuron k in the output layer; symbols y1, y2 . yk, yt denote outputs of the neurons in the output layer.2.2 Modelling by Applying Artificial Neural NetworksThe architecture and size of a neural

26、network depends on the problem complexity. The number of neurons in the input and output layers is decided by the selected input-output variables of the analysed system. The simulation experiments of neural network training and testing indicate the optimal number of hidden layers as well as the numb

27、er of neurons in these layers.The goal of neural network training is to find the functional relationship between the input patterns and target outputs. Before training ANN, all the available data are randomly divided into a training set and a test set. A training set of the input patterns and corres

28、ponding desired outputs or targets is presented to the network. The network computes its output pattern, and if there is an error - a difference between actual and desired output patterns - the weights are adjusted to reduce this error according to the learning law of training algorithm. The error f

29、unction is a useful indicator of the networks performance. The training algorithm attempts to minimise this criterion. When the value of the error function in an entire pass through all training sets, or epoch, is sufficiently small, a network is considered to have converged. Once the training phase

30、 is complete, the networks ability to generalise is tested against examples of the test set.More than a hundred different learning algorithms are available, but the most popular method is backpropagation. The backpropagation learning algorithm has two phases. First, a training input pattern is prese

31、nted to the network input layer. The network then propagates the input pattern from layer to layer until the output pattern is generated by the output layer. If this pattern is different from the desired output, an error is calculated and then propagated backwards through the network from the output

32、 layer to the input layer. The weights are modified as the error is propagated.Among the numerous artificial neural networks that have been proposed, backpropagation networks have been extremely popular for their unique learning capability Haykin (1993). 80% of practical ANN applications used the ba

33、ckpropagation neural networks. Development of construction project management effectiveness model by applying multilayer backpropagation neural networks is presented in chapter 4.3. Construction Project Management Effectiveness FactorsTraditionally, construction project management effectiveness is d

34、efined as the degree to which project goals and expectations are met. It should be viewed from respective perspectives of different project participants and the goals related to a variety of elements, including technical, financial, social and professional issues. Criteria are needed to compare the

35、goal level against the performance level. The criteria are the set of principles or standards by which judgment is made Lim et. A1. (1999). While effectiveness is measured in terms of goal attainment, there is ambiguity in determining whether a project is success or failure.Different factors are ide

36、ntified in project success studies. Ashley et. A1. (1987) conducted a pilot study within their research that, based on their analysis, established six determinants of construction project success. Jaselskis and Ashley (1991) developed a predictive discrete-choice model that focused on the project ma

37、nager, the project team, planning and controls. Pinto and Slevin (1988) determined a group of predictive critical success factors. Sanvido et al. (1992) established the four most critical success factors derived from the integrated building process model. Chua et al. (1997, 1999) distinguished betwe

38、en the critical success factors for different project objectives of budget, schedule, and quality using the analytic hierarchy process. They established 10 critical factors for each project objective. Overall, they identified 67 different success-related factors.Other measures of project success for

39、 particular group of projects were provided by Tiong (1996), Mohsini and Davidson (1996), Chan et al. (), Molenaar and Songer (), Zhang (). Cheng et al. () established a partnering framework to identify the critical success factors that can improve the productivity and performance of construction pr

40、ojects.Other studies of particular factors impact on construction project success was provided by Back and Moreau (), Mitropoulus and Tatum (), Faniran et al. (1998), Angelides (1999), Bower et al. (), Ford () and Jan et al. (). All the above mentioned studies revealed many different factors and the

41、ir qualitative impact on project success. This research, differently from the previous, focus on the functional relationships between the input factors and project outcome, analyses and enables to forecast quantitative impact of determined critical factors onto the effectiveness of construction proj

42、ect management. In this study the framework for the list of construction management effectiveness factors covering areas related to project manager, project team, project planning, organization and control was selected from the research conducted by Jaselskis and Ashley (1991). However, the actualit

43、y of each construction management factor was retested by interviewing construction management practitioners and the approach was modified according to the interviewers opinion (Table 1).4. Development of Construction Project ManagementEffectiveness Model by Applying Neural NetworksConstruction proje

44、ct management effectiveness modelling by applying neural networks consists of the following stages:selection of the variables of the construction project management effectiveness neural network model (CPMEM);selection and preparation of training data for the CPMEM;designing and training the construc

45、tion project management effectiveness neural network;evaluation of the importance of a particular input factor to the CPMEM output by applying a sensitivity analysis technique;identification of the key construction project management effectiveness factors and modification of the CPMEM;determining th

46、e validation range of the CPMEM practical applications.Construction project management effectiveness factors are the input variables of the CPMEM. The output variable of this model is the construction project management effectiveness in terms of construction cost variation. The construction project

47、cost variation was calculated by equation:Q = （PI - FI）/PI* 100%where PI - predicted construction project cost; FI - actual construction project cost. The present study is based on a set of data obtained in a questionnaire survey on construction project management effectiveness factors from construc

48、tion management organizations in Lithuania and the USA. Twelve Lithuanian companies presented information on 32 completed construction projects. The average size for the projects is 4.3 million Litas (1.6 million USD) and the mean duration is 7 months. 27 US construction management companies present

49、ed information on 54 completed construction projects with the average size of 30.1 million USD and the mean duration of 14 months. Statistical analysis proved that those two groups of the projects belong to the same statistical population. Thus, neural network model was trained with 76 project sampl

50、es and retested with 10 project samples. The construction project management effectiveness neural network model had been developed using NEURAL NETWORKS TOOLBOX by MA TLAB.A neural network works best when all its inputs and outputs vary within the range 0 and 1. Preparation of the training data and

51、statistical computations had been performed by applying Microsoft Excel. The input data - project management factors - was classified into six groups and the output data - the percentage of the construction cost variation in loss or profit - was classified into five groups (Table 2). The number of n

52、eurons in the input and output layer was decided by the number of input and output variables of the construction project management effectiveness neural network. Thus, the input layer had 27 neurons and the output layer had 5 neurons, representing five classes of the construction cost variation. The

53、 number of hidden layers was determined during the neural network training.The neural network was trained to Solve the classification task by applying resilient backpropagation learning algorithm. The network performance in this study was measured by the modified regularization error function. The i

54、nterpretation of the network output is based on the Bayesian posterior probability: the construction project cost variation belongs to the class represented by the output layer neuron of the highest output value. The classification error was calculated by equation:where Tp - actual class of project

55、cost variation; Pp- class of project cost variation predicted by neural network; p - construction project index; q - number of examples for testing.All construction management effectiveness factors were incorporated into the model at the first stage of model development. The initial network model co

56、mprised 27 neurons in the input layer with 9 neurons in the hidden layer and 5 neurons in the output layer. In order to understand the importance of a particular input to the network output, a sensitivity analysis technique was applied. The priority level for each factor was set based on their diffe

57、rent impact to the project results. Insignificant factors were trimmed from the network gradually by eliminating the least important factors, respectively to the results of sensitivity analysis. In this model development stage 12 key determining construction management effectiveness factors were ide

58、ntified. Nine key factors showed positive influence on the CPMEM output. The higher values of these factors allow improving the construction project management effectiveness. Three key factors, i.e. PM subordinates, independent constructability analysis, and control system budget, showed negative in

59、fluence on the CPMEM output. These factors appear to be associated with project complexity and risk. The higher project complexity and the higher level of risk degree means the higher values of these three factors: there are more employees and subcontractors supervised by PM, the cost of independent

60、 constructability analysis as well as control budget is respectively higher (Table 3).The final neural network model was built with 12 neurons in the input layer, 4 neurons in hidden layer and 5 neurons in the output layer.The established CPMEM represents the input-output functional relationships re