Modified Mamdani-fuzzy inference system for predicting the cost overrun of construction projects

Abstract

Cost overruns are a common worldwide problem in the construction industry; improved proactive risk management and cost control are much needed. Several models have been proposed, but all have weaknesses, particularly in data demands and the severity of critical risks or uncertainties associated with expert judgment. In response, this study develops a new 3-part model based on the Mamdani-type fuzzy inference system (FIS) to predict the cost overrun of construction projects. The first part assesses the weight of each expert, evaluating the severity of cost overrun factors. The second part contains a list of 40 in-built cost overrun factors and their degree of severity, while the third part establishes the relationships of every factor's occurrence probability and severity to predict the cost overrun of a specific project. The severity of each factor is assessed based on a survey of 31 randomly selected experts in the Saudi Arabian construction industry. The model is demonstrated on two completed projects in Saudi Arabia. For each project, this involves a group of project-based experts rating the probability of occurrence of each factor on that project and applying this to the factor severity list to obtain a predicted cost overrun (PCO) for the whole project. The model is validated for robustness by sensitivity analysis comparing the predicted and actual whole project cost overrun and shown to be of practical value in assessing critical risks and predicting the likely amount of cost overrun. The model is equally applicable in the early project stages.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 32 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.