A comparison of fuzzy logic and multiple regression analysis models in determining contingency in international construction projects

Purpose - The main objective of this paper is to show how fuzzy logic and multiple regression analysis (MRA) techniques can be used by construction companies for determining the size of contingency that will be included in bid prices for international construction projects in a more systematic way and to compare their modelling performances. Design/methodology/approach - The steps followed in the execution of this study mainly consists of: conducting a literature review on international construction in order to identify the factors that may affect contingency amounts that will be included in bid prices for international construction projects; developing the general framework of the proposed contingency estimation model; designing a questionnaire based on the information gathered from the literature review, delivering these questionnaires to construction experts, and obtaining the actual data of 36 international construction projects; developing a fuzzy logic model based on expert judgments and three multiple regression analysis models (MRAM) using the collected data; and comparing the performances of these approaches. Findings - In this study, a fuzzy logic model and three MRAM were developed. Their modelling performances were compared using actual data obtained from 36 international construction projects that had been completed by 20 large-scale Turkish construction companies in 14 different countries. It is found that the developed fuzzy logic model outperforms the MRAM built for the studied projects. Originality/value - This study shows that fuzzy logic and MRA techniques can be successfully used by construction companies, which predominantly do business in foreign countries, for estimating the size of cost contingency that will be included in bid prices for international construction projects. The modelling performances of fuzzy logic and MRAM are also compared.