Multi-Criteria Decision Support for Sustainable Supplier Evaluation in Mining SMEs: A Fuzzy Logic and TOPSIS Approach

Abstract

Background: Improving operational efficiency in the mining industry increasingly de-pends on a mature asset management framework and the careful selection of reliable, sustainable suppliers for systems, personnel, equipment, and services. Given the complexity of mining operations and the growing use of digital tools, choosing the right maintenance management system requires a robust decision-making process that considers economic, environmental, and social sustainability factors. Methods: This study develops and compares two multi-criteria decision-making approaches, a ranking method and a fuzzy logic-based model to evaluate four maintenance management systems against fifteen sustainability-related criteria. Expert opinions from executives and operational managers in the South African mining sector were gathered, focusing on factors such as cost, integration, reliability, ease of use, inventory control, and predictive capabilities. Results: The ranking method produced a clear, quantitative order of preference, while the fuzzy model addressed uncertainty and subjectivity in expert judgments. Both methods identified the same top choice: UPKEEP, followed by SAP, FIIX, and LIMBLE. Conclusions: This comparison shows that combining fuzzy logic with sustainability-focused evaluation can improve the flexibility and reliability of supplier selection in asset management. The proposed approach offers practical guidance for aligning maintenance system choices with broader sustainability goals in mining operations.

1. Introduction

The ability of mining companies, particularly small and medium-sized enterprises (SMEs), to adopt innovative and robust asset management practices is vital for maintaining competitiveness, meeting stakeholder expectations, and ensuring long-term sustainability [1]. Effective asset management requires systematic and coordinated processes for optimizing the performance, costs, and risks of assets throughout their life cycles, in alignment with strategic business objectives [2]. Within this context, the selection and evaluation of reliable suppliers for systems, equipment, and maintenance services have emerged as critical factors in enhancing operational resilience and sustainability performance [3].

However, the inherently complex operational structures of mining SMEs, combined with the unpredictable market and regulatory conditions in which they operate, introduce significant uncertainty into supplier selection and decision-making processes [4]. Traditional decision-making models often fail to address this complexity, as they may not adequately capture qualitative factors or the subjective judgments of industry experts. As a result, mining companies risk suboptimal supplier choices, which can lead to inefficiencies, higher costs, and missed opportunities for sustainability improvement.

One of the most important steps in supply chain management is supplier selection, when decision-makers choose the best supplier for the goods or services they want to buy [5]. Businesses typically search for the best suppliers to improve performance and maintain existing relationships over time. Additionally, most mining costs are attributed to the cost of resources; therefore, selecting a supplier is crucial to mine financial stability. The cost of the resources and services needed to produce a product usually makes up 70% of the selling price [6,7,8]. As a result, an economical supplier can significantly reduce the costs associated with the supply chain. Additionally, the profitability of an organization is directly impacted by its suppliers [9,10].

The first stage of the supplier selection process involves identifying and confirming the provider. Following this, the contract is signed. One of the most crucial supply chain processes is selecting a supplier, even though it may seem straightforward [11]. Therefore, selecting a supplier wisely based on the situational requirements is essential to maintaining competitiveness, profitability, and security [12,13,14]. Questions about selecting a supplier and figuring out which one is most likely to fulfil the requirements are common during the SS process. Since the process of choosing a supplier is a multi-criteria problem, the answer to this question depends on a number of qualitative and quantitative factors [15,16]. The selection of suppliers is greatly impacted by these kinds of factors [17,18,19,20]. Therefore, weighing these factors and trade-offs is necessary to choose the best supplier. However, it is not possible to compromise on all possible supplier selection criteria. Additionally, not every criterion has an equal impact on the choice of providers. The SS process may become more complicated and the likelihood of selections being made improperly increases when suppliers are selected based on meaningless criteria. Additionally, choosing a supplier based on a wide range of criteria could lead to a poor selection process, which could negatively impact the performance and profits of the company. Therefore, it is essential to choose pertinent criteria for the selection process when selecting a provider. Therefore, it would be great to recognize each important requirement from the wide range of criteria. Since this is a multi-criteria problem, it is possible to use multi-criteria decision making (MCDM) processes [21]. By applying scientific analytical methods to real-world issues, MCDM provides an evaluation framework that can assist decision-makers (DMs) in developing practical solutions [22].

This research proposes and tests a multi-criteria decision support framework that integrates both a ranking method and a fuzzy logic-based approach. This hybrid model aims to enhance supplier evaluation by systematically incorporating sustainability-related criteria and accounting for the uncertainty and subjectivity inherent in expert evaluations. By focusing on the mining SME context, this study contributes practical insights and a replicable method for aligning supplier selection decisions with broader sustainability and asset management goals.

The research further addresses the following crucial research questions:

• What are the most relevant sustainability-related criteria for evaluating suppliers in the context of mining SMEs?
• How can a hybrid decision-making framework combining a conventional ranking method and fuzzy logic address the inherent uncertainty and subjectivity in supplier evaluation?
• How do the results of the hybrid framework compare to conventional decision-making methods in selecting the most appropriate supplier?
• What practical insights can be drawn from applying this hybrid framework to a real-world mining SME case study to guide future supplier selection decisions?

This paper’s remaining sections are organized as follows. The pertinent background information is compiled in Section 2, which also offers a targeted summary of important literature on integrated approaches to supplier selection. The proposed integrated method is discussed in Section 3. In this section, the TOPSIS model and fuzzy approach used to rank suppliers are discussed in detail. To validate the suggested approach, a case study involving a mining organization is presented in Section 4. Section 5 concludes the study.

2. Literature Review

2.1. Asset Management in the Mining Sector

Asset management is the coordinated activity of an organization to realize value from an asset [23]. The mining industry is highly dependent on a mature asset management program for meeting up with production targets and smooth operations of activities. Continuous information on asset status and performance, reliable communication channels, and adequately documented professional knowledge from previous experience will be critical to the efficacy of asset management in the operation and maintenance phase [24,25,26]. Several researchers have studied the influence of industry 4.0 technologies on asset management in eradicating the complexity in the operation and maintenance phase of assets [27,28]. These technologies help optimize equipment, energy, utilities, and materials, while enhancing decision-making processes and monitoring [29].

An effective asset management program that incorporates digital technologies for real-time data collection, predictive maintenance, and efficient resource management is necessary for mining operations to run as efficiently as possible. Yet, the sector still has difficulties utilizing these technologies to their full potential, especially when it comes to successfully implementing Computerized Maintenance Management Systems (CMMSs).

Computerized Maintenance Management System (CMMS)

The management of operations and maintenance within the mining industry is inherently complex, and poor management can severely impact profitability [30]. The implementation of CMMS software has been shown to improve maintenance reliability and performance, but successful adoption remains a challenge. Studies have shown that 25–40% of CMMS implementations are successful, and only 6–15% of users fully utilize these systems [30]. Key barriers to successful implementation include the selection of an inappropriate CMMS supplier, organizational resistance to new maintenance strategies, inadequate IT infrastructure, and a lack of sufficient resources for full implementation. Operational success depends on selecting the appropriate CMMS supplier. The management of maintenance plans, inventories, expenses, and service quality, all of which have a major impact on operational excellence can be facilitated by a well-integrated CMMS. Because CMMS implementation is complicated, choosing the appropriate provider requires a methodical approach, which this study will examine in connection to mining SMEs.

It is vital to have robust tools to support the complicated process of operation and maintenance management in an effective manner, to provide information that detects key maintenance issues that impact the “iceberg’s” hidden costs, and to improve overall business performance. A CMMS or eAM software is one solution that can provide this assistance and selecting the right CMMS is a crucial part of attaining operational excellence in the mining industry.

2.2. Multi-Criteria Decision Making

Due to several decision levels and different elements to consider, the sustainable supplier selection problem is a very complicated topic [31]. The basic concept of sustainability in supplier selection is to strike a balance between environmental and social issues while pursuing economic goals [32]. Due to its impact on numerous areas such as inventory management, production planning, maintenance scheduling, financial resources, and the environment, supplier selection has long been a strategic cornerstone for the success of mining companies [33]. Furthermore, selecting a supplier is dependent on numerous factors that are often conflicting, as such, managers can benefit from a structured decision-making strategy, such as multi-criteria decision making (MCDM) methodologies [34]. MCDM can consider a variety of factors, including sustainable criteria.

Suppliers provide the components, technologies, and materials required for the supply chain system to operate. Given that procurement expenses make up between 70% and 80% of the production costs of most businesses, supplier procurement procedures can significantly affect a company’s profitability. Furthermore, these suppliers’ resources and capabilities have a significant impact on the company’s turnover [35]. Therefore, one of the primary goals of any supply chain management system is to efficiently manage the movement of funds, information, and resources to satisfy customer requests and achieve overall business goals [36,37]. Suppliers are the main operational engine that can either increase or diminish the supply chain’s effectiveness [38]. However, there are other drawbacks that complicate the supplier selection procedure. Finding appropriate and pertinent criteria to include in the evaluation and selection process is one of the primary issues with SS [39]. Objectivity, specificity, and comprehensiveness should be the guiding principles for choosing suppliers. Before selecting suppliers, businesses need to develop a comprehensive and accurate evaluation procedure. Financial capacities, equipment management, human resource development, quality control, cost control, technological development, user satisfaction, delivery agreements, and environmental awareness are among the criteria gathered from literature research [40]. To identify cold supply chain (CSC) suppliers, the authors of [41] looked at fifteen key factors. According to their findings, “utilization of resources” is the most important factor.

It is more challenging to reach an agreement when using MCDM methods since they consider preferences across a range of quantitative and qualitative criteria, which are frequently incompatible and challenging to reconcile. Among the disciplines that have an impact on the development of MCDM techniques are information systems, computer science, economics, and behavioral decision theory. Previous research has proposed a variety of MCDM methodologies to help companies select qualified suppliers. The techniques for order preference by similarity to ideal solution (TOPSIS), fuzzy set theory (FST), data envelopment analysis (DEA), the analytical hierarchy process (AHP), and multi-objective programming are a few of these methodologies [42,43]. These popular and conventional approaches have been enhanced or integrated by researchers to satisfy the requirements of the decision-making scenarios [44,45,46,47].

However, most of the research focused on ideas and methods related to supplier selection, by either neglecting the creation of criterion systems or assessing them qualitatively using existing literature or expert opinion [48]. The quality of the provider that is finally selected is greatly influenced by the quality of the decisions made in the earlier phases [49]. The number of studies on SS using traditional methodologies has significantly increased in the literature in recent years due to the complexity of supplier assessment challenges and the unpredictable nature of human reasoning [50,51,52,53]. However, machine learning can manage the SS procedure [54,55].

The Fuzzy Analytic Hierarchy Process, which, in addition to the advantages of the AHP, incorporates the ability of considering the fuzziness of the data while selecting the preferences of the many criteria given, is one of the most widely used methods [56,57]. Furthermore, the AHP, and even the ANP’s integration with other approaches are quite popular due to their flexibility, ease of understanding, and capacity to adapt to a variety of scenarios using a set of defined criteria [58,59,60]. FAHP was frequently used in conjunction with TOPSIS or Fuzzy TOPSIS. In this way, a combined strategy is employed to find the best green suppliers for an Iranian automaker [61]. DEMATEL and its fuzzy counterpart can also be supported by these MCDM approaches. In the field of supplier selection, FDEMATEL approaches are widely established [62]. An approximate FDEMATEL technique has just been proposed and implemented on a can plant in Jordan [63].

The incorporation of sustainability factors into decision-making frameworks for small and medium-sized mining firms (SMEs) is noticeably lacking, despite the substantial amount of research on supplier selection in the mining industry. Although supplier selection with a focus on sustainability has been studied in a number of industries [31,32,33], little is known about the particular difficulties faced by mining SMEs, such as operational complexity and resource limitations. Furthermore, the majority of current research relies on conventional approaches for making decisions, with little investigation of hybrid models that combine sustainability criteria and fuzzy logic.

3. Materials and Methods

3.1. TOPSIS Model

The TOPSIS model is a method which can assist in decision making by ranking several alternatives based on their similarity or proximity to an ideal solution and has been widely applied to various research domains during the past few decades [64]. In this study, the TOPSIS model is applied to preliminarily determine the performances of the CMMS alternatives with respect to the sustainability factors relevant to the mining sector, and thereafter rank the CMMS. The TOPSIS model is formulated in the following steps:

Step 1: Determination of the CMMS performance with regards to sustainability factors.

This is carried out using the questionnaires designed for this study. The performances of the CMMS were depicted on the questionnaires by the respondents using the linguistic scale in

Table 1. Linguistic scale for the TOPSIS model.

Semantic Attributes	Corresponding Values
Very high	5
High	4
Moderate	3
Low	2
Very low	1

.

Step 2: Design of the normalized decision matrix.

The normalized decision matrix can be computed as follows:

D = [d_ij]_mxn where i = 1, …, m  j = 1, …, n

(1)

where D is the normalized decision matrix with m rows and n columns.

The sustainability criteria are categorized into two types, namely, “benefit criteria” which signifies that the increase in magnitude is favorable, and “cost criteria” which signifies that the decrease in magnitude is favorable.

Benefit criteria

$$d_{ij}=\left({\displaystyle \frac{r_{ij}}{q_j^{+}}},{\displaystyle \frac{m_{ij}}{q_j^{+}}},{\displaystyle \frac{q_{ij}}{q_j^{+}}}\right)\mathrm{where}{q}_j^{+}=max{q}_{ij}$$

(2)

Cost criteria

$$d_{ij}=\left({\displaystyle \frac{r_j^{-}}{q_{ij}}},{\displaystyle \frac{r_j^{-}}{m_{ij}}},{\displaystyle \frac{r_j^{-}}{q_{ij}}}\right)\mathrm{where}{r}_j^{-}=max{r}_{ij}$$

(3)

Step 3: Computation of the weighted normalized decision matrix.

The weighted normalized decision matrix is calculated as follows:

$$W={⌊w_{ij}⌋}_{mxn}i=1,\dots ,mj=1,\dots ,n$$

(4)

$$W_{ij}=d_{ij}\times v_{ij}$$

where v_ij is the weight of the ith criterion for the jth CMMS alternative.

Step 4: Calculation of the positive ideal solution (P⁺) and negative ideal solution (P⁻).

Benefit criteria

$$P^{+}=\left(w_{ij}^{+},w_{2j}^{+},\dots ,w_{ij}^{+}\right)$$

(5)

$$P^{-}=\left(w_{ij}^{-},w_{2j}^{-},\dots ,w_{ij}^{-}\right)$$

(6)

Cost criteria

$$P^{-}=\left(w_{ij}^{-},w_{2j}^{-},\dots ,w_{ij}^{-}\right)$$

(7)

$$P^{+}=\left(w_{ij}^{+},w_{2j}^{+},\dots ,w_{ij}^{+}\right)$$

(8)

where $w_{ij}^{+}=max\left\{w_{ij}\right\}$ and $w_{ij}^{-}=min\left\{w_{ij}\right\}$ i = 1, …, m, j = 1, …, n.

Step 5: Determination of the separation measures of each CMMS alternative from the positive and negative ideal solutions. This is demonstrated as follows:

$$\begin{gathered} b_j^{+}={\sum }_{j=1}^n{b}_t\left(t_{ij},t_j^{+}\right) \\ i=1,\dots ,m \end{gathered}$$

(9)

$$\begin{gathered} b_j^{-}={\sum }_{j=1}^n{b}_t\left(t_{ij},t_j^{-}\right) \\ i=1,\dots ,m \end{gathered}$$

(10)

where b_t is the distance between two corresponding numbers on the linguistic scale.

Step 6: Computation of the closeness coefficient (A) for each CMMS alternative. This is demonstrated as follows:

$$\begin{gathered} A={\displaystyle \frac{b_j^{-}}{\left(b_j^{+}+b_j^{-}\right)}} \\ j=1,\dots ,n \end{gathered}$$

(11)

Step 7: Prioritization of the CMMS alternatives.

The CMMS alternatives are ranked with respect to the closeness coefficient with the best alternative being closest to the positive ideal solution and the worst alternative being farthest from the negative ideal solution.

3.2. FUZZY Model

3.2.1. Basic Structure of a Fuzzy Inference System

The basic structure of a fuzzy inference system consists of a fuzzification unit, a fuzzy logic-reasoning unit (process logic), a knowledge base, and a defuzzification unit. The key element of the system is the fuzzy logic-reasoning unit that contains two main types of information. The first is a data base that defines the number, labels, and types of the membership functions the fuzzy sets used as values for each system variable, and these are composed of two types: the input and the output variables. For each variable, the designer must define the corresponding fuzzy sets. The proper selection of these variables is one of the most critical steps in the design process and can dramatically affect the performance of the system. The fuzzy sets of each variable form the universe of discourse of the variable. The second is a rule base, which essentially maps fuzzy values of the inputs to fuzzy values of the outputs. This reflects the decision-making policy. The control strategy is stored in the rule base, which in fact is a collection of fuzzy control rules and typically involves weighting and combining a number of fuzzy sets resulting from the fuzzy inference process in a calculation, which gives a single crisp value for each output. The fuzzy rules incorporated in the rule base express the control relationships usually in an IF-THEN format. The if-part of the rule is called condition or premise or antecedent, and the then-part is called the consequence or action.

Usually, the actual values acquired from or sent to the system of concern are crisp, and therefore fuzzification and defuzzification operations are needed to map them to and from the fuzzy values used internally by the fuzzy inference system.

The basic system architecture for fuzzy logic is shown in Figure 1.

The system input parameters are fuzzified using the appropriate membership functions to determine the degree of membership in each input class. The fuzzified inputs are evaluated in the fuzzy inference system using a well-defined rule base formed from the data set, expert knowledge, and other sources. The result obtained is further defuzzified, leading to the output result.

3.2.2. Fuzzy Model Notation and Formulation

The notations below are applied in the fuzzy modeling process:

$x_1$ = Purchase Cost (S₁)

$x_2$ = Installation/Setup Cost (S₂)

$x_3$ = Maintenance Cost (S₃)

$x_4$ = Integration with Organization (S₄)

$x_5$ = Compatibility with other Software (S₅)

$x_6$ = Reliability and Efficiency (S₆)

$x_7$ = Interactive and User-friendliness (S₇)

$x_8$ = Report Generation (S₈)

$x_9$ = Access Control (S₉)

$x_{10}$ = Backup System (S₁₀)

$x_{11}$ = Cloud Solution (S₁₁)

$x_{12}$ = Adoption of IOT Devices (S₁₂)

$x_{13}$ = Adoption of IOT Devices (s₁₃)

$x_{14}$ = Inventory Management (S₁₄)

$x_{15}$ = Calibration Management (S₁₅)

L = Descriptor corresponding to a low category in linguistic terms

M = Descriptor corresponding to a medium category in linguistic terms

H = Descriptor corresponding to a high category in linguistic terms

$x_i$ = Corresponding input variables

$O_j$ = The developed Fuzzy rules

$Ov$ = Output value score

${FI}_{ij}$ = Fuzzy set applied to input $x_i$ in rule j

${FO}_j$ = Fuzzy set representing the output

y* = Defuzzified output

y: Universe of possible output values (e.g., CMMS measured on a scale from 0 to 100)

f = The fuzzy logic model integrates fuzzification, rule inference, and defuzzification. The output fuzzy set for rule j, denoted as ${\mu O}_j\left(y\right),$ is obtained using the max operator for aggregation, which is standard practice in Mamdani inference.

Each $x_1$ is derived from expert survey responses and is typically normalized.

Each variable $x_i$ is assessed using a qualitative scale and then transformed into predefined linguistic terms. These linguistic terms are represented by triangular membership functions. Each input $x_i$ is carried out using its corresponding membership function, as outlined in Equation System 1.

$$\mu L\left(x\right),\mu M\left(x\right),\mu H\left(x\right)\in \left[\mathrm{0,1}\right]$$

(12)

The membership function degree of input x in the fuzzy set FI is defined by Equations (13)–(15).

$${\mu }_L{x}_i:\mathrm{Membership}\mathrm{degree}\mathrm{of}\mathrm{input}{x}_i\mathrm{in}\mathrm{the}“\text{Low}”\mathrm{fuzzy}\mathrm{set}$$

(13)

$${\mu }_M{x}_i:\mathrm{Membership}\mathrm{degree}\mathrm{of}\mathrm{input}{x}_i\mathrm{in}\mathrm{the}“\text{Medium}”\mathrm{fuzzy}\mathrm{set}$$

(14)

$${\mu }_H{x}_i:\mathrm{Membership}\mathrm{degree}\mathrm{of}\mathrm{input}{x}_i\mathrm{in}\mathrm{the}“\text{High}”\mathrm{fuzzy}\mathrm{set}$$

(15)

Rule Base (If-Then Rules): Constructed based on expert-defined knowledge.

Utilizing the formulated fuzzy rules, the system of Equation (5) is derived as follows:

$$Oj:\mathrm{IF}{x}_1\mathrm{is}{FI}_{1j}\mathrm{AND}{x}_2\mathrm{is}{FI}_{2j}\dots ..\mathrm{THEN} \mathrm{y} \mathrm{is} {\mathrm{FO}}_{\text{j}}$$

(16)

The rules are evaluated using Mamdani fuzzy inference. The resulting fuzzy output for each rule is expressed as follows:

$$\mu Oj\left(y\right)=\mathrm{m}\mathrm{i}\mathrm{n}[\mu FI1j\left(x1\right),\mu FI2j\left(x2\right),\dots ..]$$

(17)

The aggregation to form a combined fuzzy output is as follows:

$$\mu Y\left(y\right)={\displaystyle \frac{max}{j}}[\mu Oj\left(y\right)]$$

(18)

This score reflects the final output value of the solution.

The fuzzy output is defuzzified into a precise output score using the following centroid method:

$$\mathrm{y}*={\displaystyle \frac{\int y.\mu Y\left(y\right)dy}{\int \mu Y\left(y\right)dy}}$$

(19)

The CMMS output solution is thus presented:

$$Ov=f(x_1,x_2,x_3,x_4,x_5,x_6,x_7\dots \dots x_{15})$$

(20)

$$Ov=Defuzzify[{\bigcup }_{j=1}^{n}Fuzzify\left(xi\right)\to Apply\left(Oj\right)]$$

(21)

4. Results

The information was gathered from different literature sources, as well as discussions with company executives and specialists, including Engineering managers, planners, supervisors, and artisans that work in the company’s maintenance department. These experts’ opinions were used to compile the data. The method is backed by a case study, which is defined as descriptive research based on information gathering, with the goal of determining the criteria for supplier selection in the South African mining sector. Emphasis was laid on the selection of computerized maintenance management systems necessary for the optimal operation of the mining industry.

4.1. Solution Procedure

Supplier Selection for the Computerized Maintenance Management System (CMMS)

The requirements that were verified necessary both practically and in terms of organizational conditions were selected from many criteria using previous research and studies, as well as surveys and interviews with managers and specialists. Then, based on professional opinion, past research, and interviews with managers, 20 criteria were given for the selection of an effective CMMS. This was summarized into six broad sections with 15 subsections.

CMMS Cost: CMMS purchase cost, CMMS installation/setup cost, and CMMS maintenance cost.

CMMS Deployment: CMMS integration with the organization’s operations and CMMS compatibility with other software.

CMMS Performance: CMMS reliability and efficiency; CMMS interactiveness and user-friendliness; inventory management and calibration management.

CMMS Data management: CMMS report generation and CMMS billing and invoicing.

CMMS Security: CMMS access control and CMMS backup system.

CMMS Industry 4.0: CMMS cloud solution and CMMS adoption of IoT devices and predictive maintenance.

The criteria for selecting a sustainable computerized maintenance management system (CMMS) are detailed in

Table 2. Selection criteria for the computerized maintenance management system (CMMS).

S/N	Variables	Parameters
1	C1	Purchase Cost
2	C2	Installation/Setup Cost
3	C3	Maintenance Cost
4	C4	Integration with Organization
5	C5	Compatibility with other Software
6	C6	Reliability and Efficiency
7	C7	Interactiveness and User-friendliness
8	C8	Inventory Management
9	C9	Calibration Management
10	C10	Report Generation
11	C11	Billing and Invoicing
12	C12	Access Control
13	C13	Backup System
14	C14	Cloud Solution
15	C15	Adoption of IOT Devices and Predictive Maintenance

. The result shows that the selection criteria for a CMMS most valued by the respondents are purchase cost, installation/set up cost, maintenance cost, integration with organization, compatibility with other software, reliability, and efficiency, interactive and user-friendliness, inventory management, calibration management, report generation, billing and invoicing, access control, backup system, cloud solution and adoption of IOT devices, and predictive maintenance. These findings are consistent with previous research since they match the economic criteria for selecting sustainable suppliers from the research provided by [64]. In essence, managers in the mining industry focus on economic criteria in the selection of a computerized maintenance management system for the operations of their organization.

4.2. Detailed Analysis of CMMS Selection Using the TOPSIS Model

Results: The CMMS performances with respect to sustainability factors relevant to the mining industry as completed on the questionnaire by industrial experts in this field are presented in

Table 3. Performances of CMMS alternatives with regards to sustainability factors.

Sustainability Factors	SAP	UPKEEP	LIMBLE	FIIX
Purchase Cost (S1)	3	4	5	4
Installation/Setup Cost (S2)	3	4	4	4
Maintenance Cost (S3)	3	4	5	5
Integration with Organization (S4)	4	5	3	5
Compatibility with other Software (S5)	4	5	3	4
Reliability and Efficiency (S6)	5	5	4	4
Interactive and User-friendliness (S7)	4	5	4	4
Report Generation (S8)	5	5	3	5
Access Control (S9)	4	5	3	4
Backup System (S10)	4	4	4	4
Cloud Solution (S11)	4	4	4	4
Adoption of IOT Devices and Predictive Maintenance (S12)	4	4	3	4
Inventory Management (S13)	4	5	3	4
Billing and Invoicing (S14)	4	5	3	4
Calibration Management (S15)	4	4	3	4

as the results of the first step of the TOPSIS process.

The normalized decision matrix shown in

Table 4. Normalized decision matrix.

Sustainability Factors	SAP	UPKEEP	LIMBLE	FIIX
Purchase Cost (S1)	0.3693	0.4924	0.6155	0.4924
Installation/Setup Cost (S2)	0.3974	0.5298	0.5298	0.5298
Maintenance Cost (S3)	0.3464	0.4619	0.5774	0.5774
Integration with Organization (S4)	0.4619	0.5774	0.3464	0.5774
Compatibility with other Software (S5)	0.4924	0.6155	0.3693	0.4924
Reliability and Efficiency (S6)	0.5522	0.5522	0.4417	0.4417
Interactive and User-friendliness (S7)	0.4682	0.5852	0.4682	0.4682
Report Generation (S8)	0.5455	0.5455	0.3273	0.5455
Access Control (S9)	0.4924	0.6155	0.3693	0.4924
Backup System (S10)	0.5000	0.5000	0.5000	0.5000
Cloud Solution (S11)	0.5000	0.5000	0.5000	0.5000
Adoption of IOT Devices and Predictive Maintenance (S12)	0.5298	0.5298	0.3974	0.5298
Inventory Management (S13)	0.4619	0.5774	0.3464	0.5774
Calibration Management (S15)	0.4924	0.6155	0.3693	0.4924

was determined by applying CMMS performances with respect to the sustainability in Equations (1)–(3) as the second step of the TOPSIS process. The third step of TOPSIS process was carried out by computing the weighted normalized decision matrix shown in Table 4 using Equation (4). Then, the positive ideal solution and negative ideal solution for the benefit criteria were determined using Equations (5) and (6), while the positive ideal solution and negative ideal solution for the cost criteria were determined using Equations (7) and (8). The separation measures from the positive ideal solution and negative ideal solution of the benefit criteria and cost criteria were computed using Equations (9) and (10) as the fifth step of the TOPSIS model.

Table 5. Weighted normalized decision matrix.

Sustainability Factors	SAP	UPKEEP	LIMBLE	FIIX
Purchase Cost (S1)	0.0246	0.0328	0.0410	0.0328
Installation/Setup Cost (S2)	0.0265	0.0353	0.0353	0.0353
Maintenance Cost (S3)	0.0231	0.0308	0.0385	0.0385
Integration with Organization (S4)	0.0308	0.0385	0.0231	0.0385
Compatibility with other Software (S5)	0.0328	0.0410	0.0246	0.0328
Reliability and Efficiency (S6)	0.0368	0.0368	0.0294	0.0294
Interactive and User-friendliness (S7)	0.0312	0.0390	0.0312	0.0312
Report Generation (S8)	0.0364	0.0364	0.0218	0.0364
Access Control (S9)	0.0328	0.0410	0.0246	0.0328
Backup System (S10)	0.0333	0.0333	0.0333	0.0333
Cloud Solution (S11)	0.0333	0.0333	0.0333	0.0333
Adoption of IOT Devices and Predictive Maintenance (S12)	0.0353	0.0353	0.0265	0.0353
Inventory Management (S13)	0.0308	0.0385	0.0231	0.0385
Calibration Management (S15)	0.0328	0.0410	0.0246	0.0328

presents the weighted normalized decision matrix derived from Table 4 for all four systems.

The closeness coefficient for each of the CMMS alternative was calculated using Equation (11) as shown in

Table 6. Closeness coefficients for CMMS alternatives.

	SAP	UPKEEP	LIMBLE	FIIX
Positive ideal solution (P+)	0.0195	0.0143	0.0485	0.0264
Negative ideal solution (P−)	0.0364	0.0435	0.0000	0.0333
(P+) + (P−)	0.0560	0.0578	0.0485	0.0598
Closeness coefficient (A)	0.6510	0.7525	0.0000	0.5576
Ranking	2	1	4	3

.

From Table 6, the closeness coefficient for the SAP system is 0.6510 and the closeness coefficient obtained for the UPKEEP system is 0.7525. Following the same approach, the closeness coefficient obtained for the LIMBLE system is 0.0000 and that of the FIIX system is 0.5576. These values are used to rank the individual systems from highest to lowest with the highest value taking the first position and the lowest value taking the fourth position. According to the results of the TOPSIS analysis implemented in the current study, the best possible CMMS for the mining industry among the alternatives is UPKEEP. This is followed by SAP, and lastly, FIIX and LIMBLE.

4.3. Detailed Analysis of CMMS Selection Using the Fuzzy Model

Based on the study, here is a structured fuzzy logic model procedure and Solution for sustainable CMMS supplier selection in mining asset management, designed to mirror the results achieved using TOPSIS (i.e., ranking CMMS alternatives as UPKEEP > SAP > FIIX > LIMBLE). This fuzzy model accounts for linguistic vagueness and human judgment uncertainty. The system architecture for the Fuzzy Solution is presented in Figure 2.

4.3.1. Fuzzy Logic Model Procedure for Sustainable Supplier Selection

The goal is to select the most sustainable Computerized Maintenance Management System (CMMS) from a set of alternatives SAP, UPKEEP, LIMBLE, FIIX based on 15 sustainability criteria, classified under economic, social, and environmental domains and compare the results obtained with the TOPSIS model. To proceed, the Fuzzy Linguistic Scale is properly defined; this helps to convert human judgments into fuzzy numbers as presented in

Table 7. Fuzzy linguistic scale.

S/N	Linguistic Term	Fuzzy Number/Triangular
1	Very Poor (VP)/VL	0.0, 0.0, 0.25/0.0, 0.1, 0.3
2	Poor (P)/L	0.0, 0.25, 0,5/0.1, 0.3, 0.5
3	Fair (F)/M	0.25, 0.5, 0.75/0.3, 0.5, 0.7
4	Good (G)/H	0.5, 0.75, 1.0/0.5, 0.7, 0.9
5	Very Good (VG)/VH	0.75, 1.0, 1.0/0.7, 0.9, 1

.

The scale is used to transform expert evaluations from the crisp matrix into fuzzy scores.

4.3.2. Construction of the Fuzzy Decision Matrix

This stage involves transforming the crisp ratings of each CMMS under the 15 sustainability criteria into their respective fuzzy equivalents.

For example:

• Purchase Cost (S1) Ratings:

  ○

  SAP = 3 → Fair → (0.25, 0.5, 0.75)

  ○

  UPKEEP = 4 → Good → (0.5, 0.75, 1.0)

  ○

  LIMBLE = 5 → Very Good → (0.75, 1.0, 1.0)

  ○

  FIIX = 4 → Good → (0.5, 0.75, 1.0)

Repeat this for all criteria and alternatives.

4.3.3. Normalization of the Fuzzy Decision Matrix

Since all criteria are benefit-type (higher is better), the fuzzy decision matrix is normalized using the following:

$${\tilde{r}}_{ij}={\displaystyle \frac{a_{ij}}{c_j^{max}}},{\displaystyle \frac{b_{ij}}{c_j^{max}}},{\displaystyle \frac{c_{ij}}{c_j^{max}}}$$

(22)

where $c_j^{max}$ is the maximum upper bound for criterion j among all alternatives.

4.3.4. Assigning Fuzzy Weights to Criteria

At this point, it is important to assign importance levels by using linguistic weights to each criterion and convert these into fuzzy weights using the linguistic scale.

As presented in

Table 8. Criterion and TFN.

S/N	Criterion	Code	Importance Level	Triangular Fuzzy Number
1	Purchase Cost	S1	High	(0.5, 0.7, 0.9)
2	Installation/Setup Cost	S2	Medium	(0.3, 0.5, 0.7)
3	Maintenance Cost	S3	High	(0.5, 0.7, 0.9)
4	Integration with Organization	S4	Very	(0.7, 0.9, 1.0)
5	Compatibility with other Software	S5	High	(0.5, 0.7, 0.9)
6	Reliability and Efficiency	S6	Very high	(0.7, 0.9, 1.0)
7	Interactive and User-friendliness	S7	Medium	(0.3, 0.5, 0.7)
8	Report Generation	S8	Medium	(0.3, 0.5, 0.7)
9	Access Control	S9	Medium	(0.3, 0.5, 0.7)
10	Backup System	S10	Low	(0.1, 0.3, 0.5)
11	Cloud Solution	S11	Medium	(0.3, 0.5, 0.7)
12	Adoption of IoT Devices and Predictive Maintenance	S12	High	(0.5, 0.7, 0.9)
13	Inventory Management	S13	High	(0.5, 0.7, 0.9)
14	Billing and Invoicing	S14	Medium	(0.3, 0.5, 0.7)
15	Calibration Management	S15	Low	(0.1, 0.3, 0.5)

, importance levels are assumed based on typical mining sector priorities as described in the table. These fuzzy weights are used to multiply with the normalized fuzzy decision matrix to compute the weighted matrix. The importance level can be adjusted based on expert elicitation or Delphi method if available.

4.3.5. Fuzzy Weighted Normalized Matrix Computation

This stage involves multiplying the normalized values by the fuzzy weights and determining the Fuzzy Positive Solutions as follows:

$${\tilde{v}}_{ij}={\tilde{r}}_{ij}\otimes {\tilde{w}}_j$$

(23)

*Determine the Fuzzy Positive Ideal Solution (FPIS) and Fuzzy Negative Ideal Solution (FNIS).

FPIS:

$${\tilde{A}}^{+}=\left(\mathit{max}{c}_{ij},\mathit{max}{b}_{ij},\mathit{max}{a}_{ij}\right)$$

(24)

FNIS:

$${\tilde{A}}^{-}=\left(\mathit{max}{c}_{ij},\mathit{max}{b}_{ij},\mathit{max}{a}_{ij}\right)$$

(25)

For each criterion, find the highest and lowest fuzzy performance values across all alternatives.

4.3.6. Calculation of the Distance from FPIS to FNIS and the Closeness Coefficient

Using the vertex method to calculate the distance of each alternative from FPIS to FNIS, the closeness coefficient (CCi) is computed. The higher the CCi values, the better the alternatives.

4.3.7. Ranking the Alternatives

The final stage is to sort the alternatives in descending order of their closeness coefficient (CCi). The alternative with the highest CCi is considered the best. The output result obtained from the fuzzy solution toolbox (FST) is presented in

Table 9. Expected output result obtained from FST.

S/N	CMMS	CCi	Rank
1	UPKEEP	Highest	1st (Best)
2	SAP	Second Highest	2nd
3	FIIX	Third highest	3rd
4	LIMBLE	Lowest	4th (Least)

.

This fuzzy logic procedure captures expert uncertainty and provides a robust decision-making framework similar to TOPSIS, while being more suitable when judgments are imprecise or linguistic in nature. The solution obtained in this study shows that the fuzzy solution obtained is quite consistent with the TOPSIS result.

5. Conclusions

This study explored the selection of sustainable and reliable suppliers for Computerized Maintenance Management Systems (CMMSs) as a critical component of matured asset management in the mining sector. Recognizing the multifaceted nature of supplier selection encompassing economic, environmental, and social criteria, a dual-model approach was employed using both the TOPSIS method and a Fuzzy Logic-Based Decision Support System to evaluate four CMMS alternatives: UPKEEP, SAP, FIIX, and LIMBLE. The TOPSIS model provided a structured, quantitative ranking based on distance from ideal and anti-ideal solutions, while the Fuzzy Logic model addressed the subjectivity and uncertainty inherent in expert judgments using linguistic variables and triangular fuzzy numbers. Despite the methodological differences, both models produced consistent rankings, identifying UPKEEP as the most suitable CMMS alternative for supporting sustainable asset management in the South African mining industry, followed by SAP, FIIX, and LIMBLE. The comparative application of these models demonstrates that integrating fuzzy logic into decision-making frameworks enhances robustness and realism, especially in contexts involving qualitative assessments and ambiguous expert input. This approach offers practical value to decision-makers in the mining industry, enabling more informed and transparent supplier selection aligned with strategic sustainability goals. Future studies may consider extending the hybrid model with additional MCDM techniques or applying it to other sectors where supplier sustainability is critical.

5.1. Managerial Implications

In terms of management, this study offers useful information to decision-makers in the mining industry. Using a hybrid model that combines fuzzy logic and TOPSIS, managers may choose suppliers more transparently and intelligently, ensuring that their decisions are in line with strategic sustainability objectives. In order to improve the overall resilience and effectiveness of asset management systems, the results show how crucial it is to take into account economic, environmental, and social sustainability criteria when evaluating suppliers. Usability, dependability, and integration capabilities are critical considerations when mining SMEs choose their suppliers, as demonstrated by the discovery that UPKEEP is the best CMMS substitute. This strategy provides a methodical and repeatable way to enhancing asset management and supplier procurement decision-making.

5.2. Future Direction

Future research might investigate expanding the hybrid model used in this study by adding more MCDM strategies or investigating the incorporation of machine learning models to improve supplier selection procedures even further. Furthermore, although this study concentrated on the mining industry in South Africa, the approach may be extended to other industries like manufacturing or energy, where choosing suppliers sustainably is crucial. The use of this concept in other cultural or economic contexts could also be studied to gain a better understanding of its universal applicability. Lastly, future research might evaluate how choosing suppliers based on sustainability affects financial results and operational effectiveness over the long run across a range of industries.