Preliminary Resource-based Estimates Combining Artificial Intelligence Approaches and Traditional Techniques


Date

2016

Publication Type

Conference Paper

ETH Bibliography

yes

Citations

Altmetric

Data

Abstract

The fate of many construction projects is determined using preliminary project cost estimates. These estimates play a key role during the conceptual phase of projects; as in many cases they are the primary element used to decide their viability. The lack of information and the high levels of uncertainty during the conceptual phase, make it infeasible to have reliable building information models that could be used to generate quantity takeoffs for preliminary cost estimates (known in the industry as 5D BIM), in line with a Level 2 BIM maturity. This paper presents a way to combine artificial intelligence (case-based reasoning and neural networks), with traditional techniques (regression analysis), to develop accurate estimates of the resources needed in a project (e.g., construction material quantities). These estimates of resources can then be coupled with unit cost information to make preliminary resource-based cost estimates. The clear division between the technical and financial components of such an estimate gives improved decision support to project managers and decision makers. This enhances the tracking and control mechanisms which could be used to check the estimates prepared in subsequent project phases. The combination of case-based reasoning with regression analysis and the use of neural networks has shown an improved performance in the estimation of the amount construction material quantities. The proposed combination was used to estimate the amount of concrete, reinforcement, and structural steel required for the construction of tall-frame structures. The results show lower errors (overall mean absolute percentage error-MAPE) for the combined models (2.55%) when compared to the regression models (12.01%), neural network models (5.84%), and case-based reasoning models (9.30%).

Publication status

published

Book title

Selected papers from Creative Construction Conference 2016

Volume

164

Pages / Article No.

261 - 268

Publisher

Elsevier

Event

5th Creative Construction Conference (CCC 2016)

Edition / version

Methods

Software

Geographic location

Date collected

Date created

Subject

Case-based reasoning; Hybrid estimation models; Neural networks; Regression analysis; preliminary estimate of resources

Organisational unit

03859 - Adey, Bryan T. / Adey, Bryan T. check_circle
02226 - NSL - Netzwerk Stadt und Landschaft / NSL - Network City and Landscape
02655 - Netzwerk Stadt u. Landschaft ARCH u BAUG / Network City and Landscape ARCH and BAUG

Notes

Funding

Related publications and datasets