**2. Literature Review**

The problem often discussed in the literature are the factors influencing cost overruns in construction projects. According to [2], the main reasons of cost overruns are the increasing cost of resources (labor, materials, machinery), changes in design specifications, land acquisition and resettlement as well as changes in currency exchange. Chen and Hu [6] identified the following main reasons of cost overruns: delay in construction period, engineering quantity increase, and lack of technical skill and experience. Cantarelli et al. [7] investigated the causes of cost overruns in construction projects and categorized them into four main explanations for cost overruns, i.e., technical, economic, psychological, and political. Specific examples of factors were identified for each of these categories. The results of research performed by Phama et al. [8] show that four factors—risks, resources, incompetence of parties, and components, transportation, and machinery cost—are important. Firm policies, project policies, and poor collaboration of parties are not very important for cost overrun. Shaikh [9] identified five main factors as common in causing time and cost overrun in megaprojects in Pakistan. These main factors are financial issues, weather conditions, political approach, design changes, and owner interference. In [10], the authors concluded that the most significant cost overrun factors are schedule delay (47%), improper planning and scheduling (47%), frequent design changes (45%), frequent changes to the scope of work (43%), and inaccurate time and cost estimates of the project (42%). In [11], the authors identified 44 factors affecting cost overrun. Of these, 11 have a decisive influence. Sohu et al. [12] identified nine major causes of cost overrun from professionals working with contractors in highway projects in Pakistan. Catalão et al. [13] presented a methodology using the existing methods but taking into account political, legal, regulatory, and economic determinants. The analysis suggests that these factors have been underestimated in the literature but are of great importance in understanding cost overruns.

Many authors draw attention to the complexity of cost overruns, emphasizing that the factors causing overrun can only be understood by looking at the whole project system in which it occurs and how several variables dynamically interact with each other [14]. The relationships between the different characteristics of the project and cost overrun were studied, for example, in [15–20]. Many authors also analyze the generating process of cost overruns along the various phases of the project life cycle [21–23].

Another extremely important issue is the possibility of predicting the risk of cost overruns and the amount of such overruns. The risk of cost overruns is dynamic, interdependent, complex, subjective, and fuzzy, especially in large and complex projects [24]. This is the reason why many researchers have

attempted to apply fuzzy set theory to solve problems related to cost overrun. Sharma and Goyal [25] proposed a fuzzy-based model to estimate the risk magnitude of the same factors influencing cost overrun. Fuzzy sets were also applied by Marzouk and Amin [26], Knight and Robinson-Fayek [27], and Plebankiewicz [5].

Ghazal and Hammad [28] proposed a Knowledge Discovery in Databases (KDD) model, which may supplement the traditional estimation methods and provide more reliable final cost forecasting to overcome the cost overrun problem. In [29], the authors developed a method for estimating the impact of project management maturity (PMM) on project performance. The proposed method uses Bayesian networks to formalize the knowledge of project management experts and to extract knowledge from a database of previous projects. The operation of the method is shown using the example of a large project in the oil and gas industry.

Other approaches used for the analysis of cost overrun problems include statistical methods, such as multiple regression analysis (MRA) [30], a regression and ANN models [31], and case-based reasoning (CBR) [32,33].
