Please use this identifier to cite or link to this item:
Title: General Regression Neural Network based time series modelling for prediction and analysis of construction equipment maintenance costs
Authors: Yip, Hon Lun
Degree: M.Phil.
Issue Date: 2014
Abstract: Construction equipment owners and equipment contractors often face the difficulties of forecasting the behaviour of maintenance cost as breakdown of equipment can come in sudden during its servicing period. This poses an uncertainty to equipment owners that future maintenance costs may have severe discrepancy with the estimated maintenance cost under the routine maintenance schedule. This uncertainty in turn adversely affects the financial management and replacement decision making for construction equipment by the owners. This study, which attempts to provide a better solution to this problem, applies a time series analysis based on General Regression Neural Networks (GRNN) model to address the modelling and prediction of construction equipment maintenance costs. The research covers modelling of both fleet maintenance cost and equipment lifecycle maintenance cost to provide a comprehensive analytical modelling framework for construction equipment maintenance cost problem. The results show that the use of time series approach based on GRNN gives a satisfactory result for maintenance cost modelling and prediction for both fleet maintenance cost and equipment lifecycle maintenance cost with some important implications derived by global sensitivity analysis based on the model. And the use of the GRNN model in optimal replacement model provides a near-optimal timing for equipment replacement.
Subjects: Construction equipment -- Costs.
Building -- Estimates.
Hong Kong Polytechnic University -- Dissertations
Pages: 97 p. : ill. (some col.) ; 30 cm.
Appears in Collections:Thesis

Show full item record

Page views

Last Week
Last month
Citations as of May 28, 2023

Google ScholarTM


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.