Application of reliability theory and risk management in sustainable agricultural construction infrastructure in high-risk regions

Sustainable agricultural production depends on the reliability of supporting construction and environmental infrastructure, including storage buildings, cold storage facilities, grain elevators, irrigation and drainage systems, and other technical facilities. Their operational condition influences energy consumption, storage losses, product quality, water efficiency and the resilience of regional production systems (Zurek et al., 2022). This issue falls within the field of sustainable and ecological civil and environmental engineering, particularly in relation to the operation, modernization and management of infrastructure facilities in regions exposed to environmental risks. Reliability theory is a rapidly developing field of science that analyzes, predicts, and ensures the correct operation of an object, process, or system for a specified period of time under defined conditions. It is based on probabilistic analysis, analyzing the probability of failure-free operation of a system or process. Probabilistic modeling allows for a better understanding of complex, real-world systems by taking into account their variability. Risk management is the process of making decisions and taking actions aimed at achieving an acceptable level of risk. Risk management methods include identifying threats, assessing the probability of their occurrence, and implementing preventative strategies. Effective risk management is based on data and statistics, which allows for reducing risk to an optimal level, as its complete elimination is usually impossible (Aven, 2016; Bracci et al., 2021; Damayanti, 2023; Sahani et al., 2024).

In contemporary practice, the assessment of agricultural production stability typically relies on mean (normalized) values, leading to a widespread reliance on deterministic calculation methods. However, long-term observation of these processes reveals that agricultural productivity is subject to fluctuations driven by stochastic variations in influential parameters. Consequently, the task of forecasting quantitative stability indicators for agricultural production is inherently probabilistic in nature. Therefore, precise calculation of these indicators necessitates the application of reliability and probability theory. These methodologies facilitate both the quantitative assessment of agricultural reliability and the effective management of system performance over a specified period (Gnedenko & Korolev, 2020; Mirtskhulava, 1974; Mirtskhulava, 1985; Ushakov, 1985). In this context, the aim of the study is to apply reliability theory and risk management to assess the stability of agricultural systems together with their supporting construction and environmental infrastructure.

Deterministic calculation techniques are widely used because current methods for evaluating the stability of agricultural production usually rely on mean (normalized) results. However, long-term observation of the processes being studied shows that stochastic variations in affecting parameters can cause oscillations in crop yields. As a result, predicting quantitative stability indicators in agriculture is a probabilistic undertaking by nature. Therefore, reliability theory and probability theory must be applied to calculate these indicators accurately. These techniques make it possible to calculate reliability levels, guaranteeing that agricultural conditions will support a business for a predetermined amount of time. The study was conducted as a case study of regional agricultural production systems in Georgia. The empirical material covered the period 2020–2024 and included regional data on agricultural output expressed by agricultural GDP. In addition, selected information on supporting agricultural construction infrastructure was considered, including grain and feed silos, cold storage facilities, storage chambers and other technical facilities used for the storage and processing of agricultural products. These data provided the basis for calculating and interpreting the reliability indicators of agricultural production systems in relation to the availability and role of supporting infrastructure.

The ability of agroecosystems to conduct their necessary tasks at predetermined volumes while preserving ecological balance under specified operating modes and conditions, guaranteeing the necessary quality over a specific period, could be conceptualized as the reliability of agricultural output. Reliability is a complex attribute that includes several interconnected ideas. Among the most important are:

• Dependability: the capacity to continue operating consistently for a predetermined amount of time.

• Durability: the ability to maintain functional integrity for a long time.

• The capacity to effectively repair and restore system functions is known as maintainability.

• Survivability: the capacity to endure external forces, both expected and unexpected, without experiencing systemic failure.

• Efficiency preservation is the ability to guarantee that production expenses and results stay within reasonable limits.

The reliability indicator for agricultural production, which comprises of numerous integrated subsystems, substantially reflects both the level of agricultural development and the specific cultivation technologies employed within these systems. It is important to note that the concept of reliability is utilized in two distinct contexts: as an inherent property of a given object or system, and as a quantitative metric for assessing the stability of agricultural output. To describe the threat of a system failing to perform its designated functions, the literature employs the term “risk” (Barlow, 2013; Fundamentals, 2024; Ostreikovsky, 2003; Willie & Kabane, 2024). Consequently, the level of risk is quantified through the reliability indicator.

Consequently, the reliability of agricultural production in risk-prone farming regions, driven by a multitude of factors, must be conceptualized as the reliability of a hierarchical structure of subsystems, each composed of various elements which, in turn, consist of specific components. To evaluate overall reliability, it is necessary to construct a logical framework and perform a preliminary reliability assessment for these subsystems and their respective constituent elements (Kechkhoshvili et al., 2024; Vartanov et al., 2025).

For an approximate estimation of agricultural production reliability, let us define the random function characterizing the decline in productivity as ∆Y₀. This function represents the reduction in yield levels relative to the potential yield or the absolute maximum recorded in a region over a sufficiently long observation period. One of the most straightforward generalized indicators of agricultural reliability suitable for analytical purposes is the relative aggregate production reliability index, P_gen, calculated according to the established formula proposed by Mirtskhulava (1974; 1985): (1) P=1−M[ΔY]Yp P = 1 - {{M[\Delta Y]} \over {{Y_p}}} where P is the reliability index of agricultural production [−], M[ΔY] represents the mathematical expectation (mean) of the decrease in agricultural output, and Y_p denotes the potential, scientifically determined capacity of agricultural production. For a practical estimation, Y_p may be approximated using the maximum volume of agricultural output recorded in the region over a sufficiently long observation period.

This metric serves as a robust tool for comparing the reliability of agricultural production across different regions. The index is bounded between 0 and 1, where P = 0 indicates a total failure of production, and P = 1 represents ideal reliability (Gnedenko & Korolev, 2020; Ushakov, 1985).

Table 1 presents the calculated reliability of agricultural production in Georgia for the period 2020–2024. These calculations are based on data from the National Statistics Office of Georgia and the National Bank of Georgia, with indices disaggregated by region (National Statistics Office of Georgia, 2026; National Bank of Georgia, 2026; World Bank, 2026; International Monetary Fund, 2026).

A correlation analysis was conducted to examine the relationship between the reliability of agricultural production (P) and the mean annual agricultural GDP (Y, expressed in million GEL). The results reveal a moderate positive correlation r_xy = 0.6191, which is expressed by the following linear regression equation: (2) P=0.7317+0.0037⋅Y P = 0.7317 + 0.0037 \cdot Y

In the regression model presented in equation (2), the variables are defined as follows: P – the reliability index of agricultural production, as defined in equation (1). This dimensionless indicator represents the stability of the agricultural output. Y – the mean annual value of the gross domestic product (GDP) originating from the agricultural sector, averaged over the five-year observation period (2020–2024), expressed in millions of Georgian Lari (GEL).

It is important to note that the calculated correlation between the reliability of agricultural production (P) and the mean annual agricultural GDP (Y) exhibits a direct and moderate-to-strong positive relationship. The relationship between the reliability of agricultural production (P) and the mean annual agricultural GDP (Y) is illustrated in Figure 1.

Fig. 1.

Relationship between agricultural production reliability (P) and mean annual agricultural GDP (Y, in million GEL). The trend line indicates a positive correlation r = 0.59 (own research)

As demonstrated, the correlation between agricultural GDP and production reliability is characterized as moderate and positive. It is important to note that the coefficient of determination (R² ≈ 0.35) indicates that approximately 35 % of the variance in agricultural reliability is attributable to variations in GDP. Consequently, the remaining 65 % of the variation is influenced by other critical factors, such as land reclamation practices, fertilizer application, scientific labor organization, and other systemic variables (Allen, 1992; Allanson et al., 1995; Broadberry, 2008; Lampkin & Padel, 1994; Lowe & Ward, 2007).

Among these systemic variables, supporting construction and environmental infrastructure plays an important role, as it affects storage conditions, product losses, energy consumption, water management and the continuity of agricultural supply chains. Table 2 contains information on the available agricultural construction infrastructure necessary for storing agricultural production, such as warehouses, grain and feed silos, storage chambers and cold stores.

According to official data from the National Statistics Office of Georgia for 2024, there are 39 grain elevators and 387 cold storage facilities operating in the country. This distribution is regional, meaning that the majority of refrigeration infrastructure is located in Shida Kartli (Eastern Georgia) (59.7 %), while Kvemo Kartli (Eastern Georgia) and Kakheti (Eastern Georgia) account for the largest share of grain elevators (25.6 % and 23.1 %, respectively). Furthermore, according to statistics, approximately 63.6 % of these facilities are fully equipped with modern technology, which is critical for maintaining the reliability of Georgia’s agricultural production chain.

The pursuit of increased productivity is therefore linked to ensuring an appropriate level and technical condition of agricultural construction infrastructure. Such infrastructure is essential for the efficient operation of agriculture, while its reliability directly affects the efficiency of agricultural product storage, the reduction of losses and the resilience of regional production systems. To link the presented infrastructure data with the reliability-based approach, selected infrastructure-related factors affecting the stability of agricultural production systems are summarized in Table 3.

As shown in Table 3, the reliability of agricultural production systems is closely related to the availability and technical condition of supporting construction and environmental infrastructure. These facilities may reduce storage losses, improve resource efficiency and strengthen the resilience of regional production systems.

This study validates the use of reliability theory and risk management to assess sustainability in agricultural production systems and their supporting construction and environmental infrastructure in high-risk regions. The findings show that reliability indicators can be used to quantify the capacity of agricultural systems to maintain production levels under environmental and operational variability.

The research confirms that probabilistic modeling provides a more accurate assessment of agricultural reliability than traditional deterministic approaches. By accounting for stochastic environmental phenomena, such as droughts, floods and soil erosion, this methodology enables the identification of critical risk factors and the assessment of their impact on production stability.

The results also indicate that the reliability of agricultural production systems depends not only on production indicators, but also on the availability and technical condition of supporting infrastructure. Storage buildings, grain and feed silos, cold storage facilities, storage chambers, slaughterhouses, and irrigation and drainage systems may influence storage efficiency, product losses, energy consumption, water management and the resilience of regional production systems. These findings are consistent with the principles of sustainable civil and environmental engineering, as they emphasize infrastructure reliability, resource efficiency and risk reduction in maintaining stable agricultural production in high-risk regions.