**1. Introduction**

The number of road construction projects is increasing dramatically every year. Although project management is being more expertly implemented, there are still problems associated with cost overruns in projects [1]. One of the factors that increases the capital output ratio for a country's economy is cost overrun. Estimating the cost of projects has always been a crucial, demanding and sophisticated challenge [2,3]. Cost estimation is a process in which the total cost of a project is predicted based on the existing information [4]. Generally, cost estimation is conducted in order to set the initial budget of a project, which will ideally produce symmetry between the initial estimation and the subsequent actual cost [1]. Cost estimation presents some difficulties, such as the initial information required, the small number of databases available for road construction project costs, the low efficiency of existing cost estimation methods and the existence of uncertainties [5].

Earned Value Management (EVM) is a tool to help with controlling the progress of a project. EVM is able to illustrate the current status of projects, as well as measuring current variances [6]. To assess the progress of projects, EVM exploits three constraints: time, scope and cost. Moreover, EVM is able to predict the future parameters of projects, including the final cost, based on existing data [7–9]. This comprehensive management approach has been widely used in numerous studies and in different fields [10–14].

Artificial Neural Networks (ANNs) are an effective tool that imitates the human mind for application in various problems [15]. The first application of ANNs in construction activities took place in the late 1980s [16]. Adeli (2001) published the first scientific article regarding the use of ANNs in the construction industry [17]. ANNs are widely used in various stages of a project, including design, construction, maintenance, renovation and destruction [18]. Some examples of the use of ANNs are presented in the following.

Albino and Garavelli (1998) applied a neural network in order to rank subcontractors in construction firms [19]. Leung et al. (2001) exploited ANNs to predict the hoisting times of tower cranes [20]. Cheung et al. (2006) forecasted the performance of projects using neural networks [21]. Vouk et al. (2011) analyzed the economy of wastewater systems using neural networks [22]. Mucenski et al. (2013) estimated the recycling capacity of multistorey buildings using ANNs [23]. Chaphalkar et al. (2015) used a multilayer perceptron neural network in order to forecast the outcome of construction dispute claims [24]. Golanaraghi et al. (2019) predicted formwork labor productivity using an ANN [25]. Tijanic et al. (2019) used an ANN in order to predict costs in road construction [26]. Readers are referred to References [27–36] for further uses of ANNs for various applications in the construction industry, as well as in other fields of science.

Cost, time and quality are the three components of success in a construction project. In other words, a project in which construction is finished within the predicted cost, to the required quality and within the forecasted time can be called a successful project [37]. The cost of construction projects usually deviates from the initial estimation due to a variety of factors [38]. In other words, the costs in construction projects do not usually remain the same as they were predicted to be before the construction phase. Cost increases are normal, as can be seen in most projects [39]. According to the available literature, not many projects are finished within the forecasted cost. A lot of construction projects face both delays and cost overruns [40]. Flyvbejerg et al. illustrated that cost underestimation happens dramatically more frequently than cost overestimation [41]. Iran is a developing country, and cost overruns are common in such countries. For instance, Heravi and Mohammadian (2019) investigated 72 construction projects in Iran based on both their documentation and their actual performance. They concluded that larger projects faced higher cost overruns and delays [42]. Although EVM is able to illustrate the degree to which delays and cost shortages exist in a project on the basis of the project's previous data, it cannot provide an accurate prediction of the future status of the project [8,9].

EVM results are obtained during and after the implementation phase. Thus, having the ability to predict the future situation of the project during the implementation phase could be very useful for project managers. The novelty of this study is in using an ANN, a tool that possesses the ability to learn from existing data in order to effectively predict the future status, in order to obtain more precise future predictions [25]. In this way, hazardous situations are less likely to happen, as they will have been forecasted before their occurrence. There are few previous research studies that have attempted to address the deficiency of the earned value management system in accurately predicting a project's future status. Moreover, as mentioned before, construction projects usually face time and cost overruns, making it a permanent issue for all project managers [37]. For instance, Moura et al. conducted a research study and concluded that construction projects experienced cost overruns of 20.4% to 44.7% in comparison to the initial cost estimation [43]. Thus, the significance of this study is in enabling project managers to use ANNs instead of the traditional EVM method in order to predict a project's future status more accurately and to fill the mentioned gaps in the body of knowledge. In the current study, we chose to investigate road construction projects in Fars Province, Iran, as a case study. The findings of this study will help road construction industry members to predict cost indices more precisely in their projects.
