No full-text available
To read the full-text of this research,
you can request a copy directly from the authors.
Evaluation of models to predict the construction material quantities of cylindrical storage structures at an early project phase
Abstract
In the determination of whether or not a manufacturing plant should be constructed, it is necessary, during the early stages of a project, to make accurate estimates of how much the plant will cost. Most predictive models used in the preliminary estimation of plant costs have been focused directly on the costs, and therefore, have grouped together uncertainties related to the amount of construction materials used in construction, the construction processes employed, and the actual prices paid for these materials and the execution of these processes. In order to improve upon the accuracy of the prediction of plant costs at an early phase, the first uncertainty to be reduced is that associated with the construction material quantities (CMQs).
A study including the as-built bill of quantities from 53 cylindrical storage structures (CSSs) of different types from 13 manufacturing plants around the world has been conducted to evaluate the performance of different parametric construction-material-quantity models using different techniques.
Six models based on regression analysis, developed using different techniques, and one model based on Neural Networks, developed using Generalized Reduced Gradient (GRG) nonlinear optimization, were investigated. The models were constructed taking into consideration 13 potential independent variables (8 continuous and 5 categorical). The ability of the models to predict the quantity of concrete and reinforcement steel in CSSs was evaluated by comparison of the adjusted R2, the standard error of estimate, the mean absolute percentage error (MAPE) and plots of the cumulative probability vs. the percentage error.
It was found that the best models for the prediction of the amount of concrete (m3) and the amount of reinforcement (tons) to be used in CSS construction were built using Neural Networks. The best regression models were the ones built using the backward elimination multiple regression technique. All the models created using different techniques met the expected accuracy ranges for Class 4 and Class 5 estimates proposed by the Association for the Advancement of Cost Engineering International.
Preliminary cost estimates are the first thoughtful efforts to predict the cost of a project. These estimates heavily influence the fate of a project as they are crucial during the initial decision-making process. A great array of preliminary cost estimation methods has been developed for the different types of construction industries. These estimates typically concentrate only on project costs (often a single monetary value). Little attention has been given to the development of models for estimating the construction material quantities (CMQs) needed for the different elements that comprise a project. If the required CMQs can be accurately estimated, current unit costs can be included to determine the project cost by creating a preliminary estimate with a clear separation between technical estimates (quantities) and market fluctuations in prices (cost of materials and labor).
The objective of this doctoral thesis has been to develop a methodology that would allow project estimators to make accurate preliminary estimates of the CMQs used in construction projects in a consistent and systematic manner.
This project has confirmed that research efforts have to be made regarding the efficient procedure of
the operational maintenance and repair. Data available in the operating organizations is still very
heterogeneous and with huge differences in scale. At the same time, there is an increasing desire to
efficiently use resources and, therefore, to compare with other operating organizations.
Das Projekt hat bestätigt, dass in Bezug auf die effiziente Durchführung des betrieblichen Unterhalts
und der Instandhaltung Forschungsanstrengungen unternommen werden sollen. Die Datenlage bei
den Betreiberorganisationen ist noch sehr heterogen und in Bezug auf den Umfang nur sehr
unterschiedlich vorhanden. Umgekehrt steigt der Wille, die Mittel effizient einzusetzen und sich
hierzu auch mit anderen Betreiberorganisationen zu vergleichen.
This paper describes the development of linear regression models to predict the construction cost of buildings, based on 286 sets of data collected in the United Kingdom. Raw cost is rejected as a suitable dependent variable and models are developed for cost/m(2), log of cost, and log of cost/m(2). Both forward and backward stepwise analyses were performed, giving a total of six models. Forty-one potential independent variables were identified. Five variables appeared in each of the six models: gross internal floor area (GIFA), function, duration, mechanical installations, and piling, suggesting that they are the key linear cost drivers in the data. The best regression model is the log of cost backward model which gives an R-2 of 0.661 and a mean absolute percentage error (MAPE) of 19.3%; these results compare favorably with past research which has shown that traditional methods of cost estimation have values of MAPE typically in the order of 25%.
Pavement surface cracking has long been considered an important criterion for maintenance intervention because of its detrimental effects on pavement performance. Once initiated, cracking increases in severity and extent and allows water to penetrate the pavement. The water weakens the unbound layers and consequently accelerates the rate of pavement deterioration. Cracking prediction and its control are thus key components in determining the timing and cost of pavement maintenance. A neural network–based model is presented in this paper for predicting flexible pavement cracking. One-, two-, and three-hidden-layer backpropagation neural network (BPNN) topologies are investigated and their cracking-prediction performances compared. Based on the analysis, it is concluded that for the same optimal number of processing elements, a one-hidden-layer BPNN topology may be sufficient in achieving satisfactory results in cracking prediction; increasing the number of layers may not add any significant benefit to the performance of the model.
The activation function used to transform the activation level of a unit (neuron) into an output signal. There are a number of common activation functions in use with artificial neural networks (ANN). The most common choice of activation functions for multi layered perceptron (MLP) is used as transfer functions in research and engineering. Among the reasons for this popularity are its boundedness in the unit interval, the function's and its derivative's fast computability, and a number of amenable mathematical properties in the realm of approximation theory. However, considering the huge variety of problem domains MLP is applied in, it is intriguing to suspect that specific problems call for single or a set of specific activation functions. The aim of this study is to analyze the performance of generalized MLP architectures which has back-propagation algorithm using various different activation functions for the neurons of hidden and output layers. For experimental comparisons, Bi-polar sigmoid, Uni-polar sigmoid, Tanh, Conic Section, and Radial Bases Function (RBF) were used.
The preliminary cost estimate heavily influences the fate of a transportation project, yet it can be up to an order of magnitude off the final bid amount. Poor prediction of costs in state departments of transportation can lead to less-than-optimal project selection at the front end and delays later when funding is not adequate to cover planned projects. A demonstration is made of the potential to separate quantity uncertainty from price uncertainty. If item quantities can be predicted early, then readily available unit prices can be applied to create a semidetailed preliminary estimate. Compared with the typical practice of applying a gross cost per lane mile, the proposed approach provides a more detailed basis for tracking the effects of changes during project development. This methodology is being tested for implementation by the Texas Department of Transportation.
Cost estimating is a computational process that attempts to predict the final cost of a future project even though not all of the parameters are known when the cost estimate is prepared. Artificial neural networks are a good tool to model nonlinear systems, but the learning speed of a network is often unacceptably slow and the generalization capability is often unsatisfactorily low in solving highly nonlinear function mapping problems. In this paper, a novel neural network architecture, the logarithm-neuron network (LNN), is proposed and examined for its efficiency and accuracy in quantity estimating of steel and RC buildings. The architecture of the LNN is the same as that of the standard back-propagation neural network (BPN), but logarithm neurons are added to the input layer and output layer of the network. The results indicate that the logarithm neurons in the network provide an enhanced network architecture to improve significantly the performance of these networks in quantity estimating for buildings.
This paper uses a neural network (NN) approach to effectively manage construction cost data and develop a parametric cost-estimating model for highway projects. Eighteen actual cases of highway projects constructed in Newfoundland, Canada, have been used as the source of cost data. Rather than using black-box NN software, a simple NN simulation has been developed in a spreadsheet format that is customary to many construction practitioners. As an alternative to NN training, two techniques were used to determine network weights: (1) simplex optimization; and (2) genetic algorithms (GAs). Accordingly, the weights that produced the best cost prediction for the historical cases were used to find the optimum NN. To facilitate the use of this NN on new projects, a user-friendly interface was developed using spreadsheet macros to simplify user input and automate cost prediction. For practicality, sensitivity analysis and adaptation modules have also been incorporated to account for project uncertainty and to reoptimize the model on new historical data. Details regarding model development and capabilities have been discussed in an attempt to encourage practitioners to benefit from the NN technique.
An interactive, computer-based cost model has been developed for approximate cost estimation of reinforced concrete beam and slab construction in high-rise commercial buildings, A number of design variables, such as different structural schemes, grid sizes, number of stories, grades of concrete, grid locations, and section of beams, have been incorporated in the model to assess their effect on cost and quantities of constituents of concrete construction. The use of the model is recommended for comparative cost estimation to determine the effect of design parameters on structural cost; for approximate structural-cost estimation of an overall project; for checking of estimates for structural works; for calculation of quantity index for structural schemes and system; for budgeting of materials, and, finally, for use in various studies in building economics.
The basic assumptions of regression analysis are recalled with special reference to the use of a logarithmic transformation. The limitations imposed on inference-making by failure to comply with these assumptions are discussed and ways to avoid the limitations indicated. A systematic bias of the order of 10 to 20% which is inherent in most, if not all, prior uses of the logarithmic equation to estimate plant biomass is noted as is the correction for the bias.
Normal 0 false false falseKEY BENEFIT: Elementary Statistics Using Excel, Fourth Edition, offers a complete introduction to basic statistics, featuring extensive instruction on the use of Excel spreadsheets for data analysis. Extensive Excel instructions are provided along with typical displays of results, as well as information about Excel's limitations and alternative approaches to problem-solving. Real data in many examples help readers see the prevalence of statistics in the real world. KEY TOPICS: Introduction to Statistics; Summarizing and Graphing Data; Statistics for Describing, Exploring, and Comparing Data; Probability; Probability Distributions; Normal Probability Distributions; Estimates and Sample Sizes; Hypothesis Testing; Inferences from Two Samples; Correlation and Regression; Multinomial Experiments and Contingency Tables; Analysis of Variance; Nonparametric Statistics; Statistical Process Control; Projects, Procedures, Perspectives MARKET: for all readers interested in statistics
In the past few years, neural networks have emerged as a problem-solving technique with capabilities suited to many civil engineering problems. Among the various neural network paradigms available, back-propagation is by far the most utilized for its relatively simple mathematical proofs and good generalization capabilities. Despite its capabilities, back-propagation suffers from several problems that hinder the development of practical neural network applications. These include slow training, ill-defined knowledge representation and problem structuring, and nonguided design of an optimal network configuration for adequate generalization. This paper represents an effort to guide the process of developing practical neural network applications using back-propagation. The paper starts with a brief description of back-propagation mathematics. Some of the heuristics and techniques used to overcome back-propagation problems, particularly lack of generalization, are identified and outlined, along with areas of potential improvements to the paradigm. An application development methodology is proposed utilizing the identified heuristics and techniques. The methodology provides a structured framework for designing and implementing practical neural network applications with less effort.
It has been known for some time (DJ Finney, J. Roy. Stat. Soc. Suppl. 7:155–161, 1941) that transformation of an arithmetic data set to logarithms results in biased estimates when predicted values from a leastsquares regression are detransformed back to arithmetic units. Predicted values are estimates of the geometric mean of the dependent variable at that value of the independent variable, rather than the arithmetic mean. Since the geometric mean is always less than the arithmetic mean, detransformed predictions will underestimate the value in question. This bias affects the interpretations of allometric equations used for estimation, such as predicting fossil body mass from skeletal dimensions, and applications of allometry as a “criterion of subtraction,” in which residual variation is evaluated.
A number of parametric and nonparametric corrections for transformation bias have been developed. Although this problem is relatively unexplored in mammalian morphometrics, it has received considerable attention in other disciplines that use power functions structurally identical to the allometric equation. Insights into transformation bias and the use of correction terms from economics, limnology, forestry, and hydrology are reviewed and interpreted for application to mammalian allometry.
Adequate estimation of construction costs is a key factor in construction projects. This paper examines the performance of three cost estimation models. The examinations are based on multiple regression analysis (MRA), neural networks (NNs), and case-based reasoning (CBR) of the data of 530 historical costs. Although the best NN estimating model gave more accurate estimating results than either the MRA or the CBR estimating models, the CBR estimating model performed better than the NN estimating model with respect to long-term use, available information from result, and time versus accuracy tradeoffs.
Despite the small number of applications to date, neural networks are likely to be able to contribute to decision support in selected fields of R&D management. We identify the potential of neural networks in the application domain and compare it to `classical' applications, such as the recognition of hand-written characters. Typical neural network architectures for R&D management tend to be simple, having low complexity, and only a small number of training samples are generally available. As an example, we carry out experiments for a typical R&D management application where neural networks have to estimate the final cost of a new product under development. It turns out that neural networks based on the standard backpropagation learning algorithm perform reasonably well when the ratio between highest and lowest cost is small, even for relatively small training set sizes. Otherwise the learning algorithm tends to undervalue low cost levels, so that deviations between estimated cost and real cost are intolerably high. Future research will have to investigate a modification of the error definition of the backpropagation algorithm. Finally, a number of general statements are derived from our experience, and examples are provided where neural networks are appropriate or inappropriate in the domain of R&D management.
The determination of sample size is a common task for many organizational researchers. Inappropriate, inadequate, or excessive sample sizes continue to influence the quality and accuracy of research. The procedures for determining sample size for continuous and categorical variables using Cochran's (1977) formulas are described. A discussion and illustration of sample size formulas, including the formula for adjusting the sample size for smaller populations, is included
Factors affecting construction project cost include project-specific factors and those reflecting the characteristics of the project team. Multiple regression is often used to estimate a project's cost, but independent variables with a high degree of correlation are likely be left out of such a model. As a result, only a limited number of factors are included in the estimate of project cost and predictions from such models will not be accurate. To overcome this technical inefficiency, the aims of this study are: to identify factors that contribute to project cost, to construct a predictive project cost model using the principal component technique and to assess the relative importance of determining factors. The data are obtained from a random sample survey comprised of Singapore building projects completed after 1992 costing more than US$5 million in value. Three main groups of variables are identified, pertaining to characteristics of the project, contractors and owner/consultants. Special project requirements such as high technological level; contractor's specialized skills; and public administered contract have significant effects on cost. Other factors include contractor's technical expertise; owner's level of construction sophistication and contractor's financial management ability. The model assesses the impact of individual factors on project cost and provides a decision support tool to estimate cost more accurately.
Statistics: an introduction using R. West Sussex IBM SPSS Statistics 19 Made Simple Print Harrell, F. E. Regression modeling strategies: With applications to linear models, logistic regression, and survival analysis
- M J C D Crawley
- P R Kinnear
Crawley, M. J. Statistics: an introduction using R. West Sussex, England: John Wiley and Sons. 2005. Print
Gray, C. D., P. R. Kinnear. IBM SPSS Statistics 19 Made Simple. Hove: Phycology Press/Taylor & Francis Group. 2012. Print
Harrell, F. E. Regression modeling strategies: With applications to linear models, logistic regression, and survival analysis, New
York: Springer-Verlag. 2001. Print