Ian Flood

Towards the next generation of artificial neural networks for civil engineering

The purpose of this paper is to stimulate interest within the civil engineering research community for developing the next generation of applied artificial neural networks. In particular, it identifies what the next generation of these devices needs to achieve, and provides direction in terms of how their development may proceed. An analysis of the current situation indicates that progress in the development of artificial neural network applications has largely stagnated. Suggestions are made for advancing the field to the next level of sophistication and application, using genetic algorithms and related techniques. It is shown that this approach will require the design of some very sophisticated genetic coding mechanisms in order to develop the required higher-order network structures, and will utilize development mechanisms observed in nature such as growth, self-organization, and multi-stage objective functions. The capabilities of such an approach and the way in which they can be achieved are explored with reference to the problems of: (a) determining truck attributes from the strain envelopes they induce in structural members when crossing a bridge, and; (b) developing a decision support system for dynamic control of industrialized manufacturing of houses.

Rapid analysis of externally reinforced concrete beams using neural networks

Abstract The load-carrying capacity of reinforced concrete beams can be compromised by concrete cracking, and the intrusion of moisture, oxygen and salt that cause corrosion of the steel reinforcement. A relatively inexpensive method of repairing such beams is to bond fiber-reinforced composites to the tensile and shear faces of the beams. Unfortunately, the numeric tools used to analyze such beams (notably, finite element analysis (FEM)) are computationally expensive making them slow to arrive at an answer, especially when dealing with complicated three-dimensional composite forms. An empirical solution is therefore proposed that involves the development of a neural network model of the performance of externally reinforced beams, developed from laboratory observations of actual beam behavior.

Queuing network analysis for waterways with artificial neural networks

An Artificial Neural Network (ANN) model has been developed for analyzing traffic in an inland waterway network. The main purpose of this paper is to determine how well such a relatively fast model for analyzing a queuing network could substitute for far more expensive simulation. Its substitutability for simulation is judged by relative discrepancies in predicting tow delays between the ANN and simulation models. This model is developed by integrating five distinct ANN submodels that predict tow headway variances at (1) merge points, (2) branching (i.e., diverging) points, (3) lock exits, and (4) link outflow points (e.g., at ports, junctions, or lock entrances), plus (5) tow queuing delays at locks. Preliminary results are shown for those five submodels and for the integrated network analysis model. Eventually, such a network analyzer should be useful for designing, selecting, sequencing, and scheduling lock improvement projects, for controlling lock operations, for system maintenance planning, and for other applications where many combinations of network characteristics must be evaluated. More generally, this method of decomposing complex queuing networks into elements that can be analyzed with ANNs and then recombined provides a promising approach for analyzing other queuing networks (e.g., in transportation, communication, computing, and production systems).

Empirical Modeling Methodologies for Construction

The paper provides a review of empirical modeling and its application within construction engineering and management. The scope of application and trends in use of this approach are first assessed, and the potential for its further development is identified. This is followed by an examination of the key components of empirical modeling, namely: the structure and operation of the model and the scheme used in its development. The paper then provides a rigorous methodology that must be followed to ensure the validity and value of the end model, covering the steps: strategizing; data collation and assessment; model development; model evaluation and final selection; final validation; and implementation. The methodology is designed to cater for all forms of empirical modeling including the procedurally more demanding development algorithms that have become available in recent years, such as simulated evolution. Overall, the paper is designed to provide researchers embarking on an empirical modeling study with an overview of when it is appropriate to use this approach, what type of system to adopt, and how to ensure development of a successful end product.

Prediction of Financial Contingency for Asphalt Resurfacing Projects using Artificial Neural Networks

Historically, actual construction costs have tended to exceed initial cost estimates and budgets. Often this discrepancy is significant enough to cause problems such as depletions of budgets, disputes, and reductions in work quality. Cost contingency is an important element included in the base cost estimate to protect construction participants including owners, contractors, and architects from the risks associated with underestimating project cost estimates and overrunning cost budgets. Typically, project participants have simply calculated contingency as a fixed percentage of project cost in spite of the importance of contingency. The uniform application of this deterministic method to calculate contingency on the basis of project costs only is not appropriate for all construction projects. This paper identifies factors that influence contingency and proposes a new method for predicting the owner’s financial contingency on transportation construction projects using an artificial neural network (ANN)–bas...

Modeling blast wave propagation using artificial neural network methods

The paper reports on work concerned with the development of artificial neural network approaches to modeling the propagation of bomb blast waves in a built-up environment. A review of current methods of modeling blast wave propagation identifies a need for a modeling system that is both fast and versatile in its scope of application. This is followed by a description of a preliminary study that used artificial neural networks to estimate peak pressures on buildings protected by simple blast barriers, using data generated from, first, an existing empirical model and, second, miniature bomb-barrier-building experiments. The first of these studies demonstrates the viability of the approach in terms of producing accurate results very rapidly. However, the study using data from live miniature bomb-barrier-building experiments was inconclusive due to a poor distribution of the sample data. The paper then describes on-going research refining this artificial neural network approach to allow the modeling of the time-wise progress of the blast wave over the surfaces of critical structures, facilitating a three-dimensional visualization of the problem. Finally, the paper outlines a proposed novel method of modeling blast wave propagation that uses a coarse-grain simulation approach combined with artificial neural networks, which has the goal of extending modeling to complicated geometries while maintaining rapid processing.

Next Generation Artificial Neural Networks for Civil Engineering

At first glance, artificial neural networks appear to be one of the great success stories in the history of computing in civil engineering. In the Journal of Computing in Civil Engineering, for example, 54 out of 445 papers published since 1995 12% have used the term “neural” in their title ASCE 2006 , while the distribution of these publications by year indicates that there has been no decline in interest over the last decade see Fig. 1 . Moreover, according to the ISI Web of Knowledge Thompson Corporation, 2006 and summarized in Table 1, the two most frequently cited articles from all issues of the ASCE Journal of Computing in Civil Engineering are on artificial neural networks. The enthusiasm with which the research community has adopted this technology over the last 15 years, reporting successful applications within every branch of civil engineering, makes it difficult to ignore. Yet, a more in-depth analysis concludes that progress in applied artificial neural networks largely stagnated following the initial applications of the early 1990s. These first applications were mostly simple function models and pattern classifiers that mapped directly from an input vector to an output vector, the types of problem that have otherwise been solved using methods such as multi-variate regression analysis. Although many new applications have been found in the ensuing years along with several refinements to the technique , these have still been predominantly simple vector mapping problems. This is a far cry from the potential of artificial neural networks anticipated by many, to provide a computational device that can emulate higherlevel cognitive processes. Such a capability would allow a wealth of new problems to be tackled in civil engineering that have so far eluded solution, including, for example: determining legal compliance of designs from drawings and specifications; identifying constructability problems from the design of a building; and measuring construction progress from site images. This lack of progress in the development of artificial neural networks is also apparent when a comparison is made with the most popular computational model, the general purpose electronic digital computer. This is significant given that the initial development of artificial neural networks dates back to the mid 1950s Rosenblatt 1958 making the technology just a decade or so younger than that of electronic digital computing. Since its inception, the electronic digital computer has evolved steadily from a device comprising just a few hundred primary processing units transistors into one comprising billions organized into a sophisticated structure of higher-order functional subsystems. Artificial neural networks on the other hand, have failed to advance beyond simple applications that require rarely more than a few hundred primary processing units neurons in this case arranged with almost no higher-order structuring.

SIMULATION AND OPTIMIZATION OF LINEAR CONSTRUCTION PROJECTS

Linear construction projects include both discrete linear projects and continuous linear projects. Almost all of the existing simulation models for linear projects are based on the discrete simulation technology and cannot satisfy the modeling requirements of continuous linear projects. There is also a lack of effective optimization tools for continuous linear projects. The available optimization models are either unable to handle complex systems or are formulated only for discrete linear projects. This paper proposes an integrated simulation-GA (genetic algorithm) approach for better planning and scheduling of continuous linear projects. The combined simulation technology will be employed to accurately model different types of activities and relationships. To calculate the production rate of a continuous activity at any point in time , an equation was derived to account for the learning curve effect, and the chaotic function will be used to simulate the uncertainties and the correlations in a time series of points. Decision variables involved in the optimization problem include the number of crews, crew sizes and formations, start times, construction sequences, and slowdown/break rules. A GA model will be developed to help the project manager search for optimal decisions. The solutions generated by the GA will be evaluated though the simulation model.

Construction simulation using parallel computing environments

The paper describes and compares alternative approaches to implementing construction simulation models within a multiprocessor computing environment. Both parallel-algorithmic and neural network based methods of simulating construction processes are considered, and compared with the conventional serial-algorithmic approach. The lines along which a simulation algorithm can be divided into tasks for parallel execution on a multiprocessor are first discussed, and the merits of each approach are identified. This is followed by a brief discourse on neural networks, their application to construction simulation, and the way in which such an implementation can be implemented within a multiprocessing environment. The merits and demerits of all approaches are discussed with particular reference to a model of an excavation system. A case study comparing the speed at which each implementation can process a simulation shows the neural approach to operate approximately two orders of magnitude faster than the alternatives. The paper concludes with an indication of future research to be conducted in this field.

Construction project control using artificial neural networks

Artificial neural networks are finding wide application to a variety of problems in civil engineering. This paper describes how artificial neural networks can be applied in the area of construction project control. A project control system capable of predicting and monitoring project performance (e.g., cost variance and schedule variance) based on observations made from the project environment is described. This project control system has five neural network modules that allow a project manager to automatically generate revised project plans at regular intervals during the progress of the project. These five modules are similar in design and implementation. Therefore, this paper will present the main issues involved in the development of one of these five neural network modules, that is, the module for identifying schedule variance. A description of a graphical user interface integrating the neural network modules developed with project management software, and a discussion on the power and limitations of the overall system conclude the paper.