Next Article in Journal
The Influences of Vacuum–Surcharge Preloading on Pore Water Pressure and the Settlement of a Soft Foundation
Previous Article in Journal
A Framework to Design and Evaluate Green Contract Mechanisms for Forestry Supply Chains
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
Sustainability
Volume 15
Issue 9
10.3390/su15097667
sustainability-logo
Submit to this Journal Review for this Journal Propose a Special Issue
► Article Menu

Article Menu

share Share announcement Help format_quote Cite question_answer Discuss in SciProfiles thumb_up
...
Endorse textsms
...
Comment
first_page
settings
Order Article Reprints
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Predictive and Prescriptive Analytics in Identifying Opportunities for Improving Sustainable Manufacturing

by
Marcin Relich
Faculty of Economics and Management, University of Zielona Gora, 65-417 Zielona Gora, Poland
Sustainability 2023, 15(9), 7667; https://doi.org/10.3390/su15097667
Submission received: 30 March 2023 / Revised: 26 April 2023 / Accepted: 3 May 2023 / Published: 7 May 2023
Browse Figures
Versions Notes

Abstract

:
Environmental issues and sustainability performance are more and more significant in today’s business world. A growing number of manufacturing companies are searching for changes to improve their sustainability in the areas of products and manufacturing processes. These changes should be introduced in the design process and affect the whole product life cycle. This paper is concerned with developing a method based on predictive and prescriptive analytics to identify opportunities for increasing sustainable manufacturing through changes incorporated at the product design stage. Predictive analytics uses parametric models obtained from regression analysis and artificial neural networks in order to predict sustainability performance. In turn, prescriptive analytics refers to the identification of opportunities for improving sustainability performance in manufacturing, and it is based on a constraint programming implemented within a constraint satisfaction problem (CSP). The specification of sustainability performance in terms of a CSP provides a pertinent framework for identifying all admissible solutions (if there are any) of the considered problem. The identified opportunities for improving sustainability performance are dedicated to specialists in product development, and aim to reduce both resources used in manufacturing and negative effects on the environment. The applicability of the proposed method is illustrated through reducing the number of defective products in manufacturing.

1. Introduction

Nowadays, manufacturing companies are paying increasing attention to the permanent improvement of sustainability performance. This results not only from environmental regulations, but also from customer awareness related to ecology issues and the social responsibility of business. Sustainable product development and manufacturing are beneficial to companies from the corporate image point of view, and they have also financial benefits. For example, the reduction of energy and material consumption during production processes results in decreasing the unit production cost. Hence, manufacturing companies are more and more interested in introducing changes toward the improvement of sustainability performance. From the product sustainability perspective, these changes can be made during product design, aiming at eco-friendly products.
The design process is one of the most important processes engaged in by manufacturing companies, dedicated to improving customer satisfaction and reducing the negative impacts of a product on the environment. Changes incorporated into product design can improve sustainability performance in different stages related to the product life cycle. For instance, better quality of materials used in manufacturing can reduce the number of defective products, as well as increase a product’s reliability, the period of its usage, and the recyclable properties in its post-use stage.
Today, the majority of manufacturing companies are reflecting on their business processes, including design and production processes, in databases of their information systems. The design process is supported by computer-aided design and engineering-aided design software. In turn, manufacturing processes can be supported by enterprise resource planning or computer-aided manufacturing systems. Sales and marketing processes can also be supported by a customer relationship management system through registering customer complaints, customer requirements, and customer feedback regarding advertising campaigns. As a result, company databases can be used to retrieve information that can be used by the research and development (R&D) department and production department to improve the sustainability performance of the business. Data-driven approaches and prescriptive analytics seem to be suitable tools for assessing sustainability performance, recognizing patterns among data, and, finally, performing simulations toward increasing sustainability performance.
Business analytics often refers to three areas: descriptive, predictive, and prescriptive analytics [1,2]. Descriptive analytics of sustainable manufacturing includes mainly standard reports and ad hoc reports providing information about sustainability performance in the various areas of manufacturing. These reports can be based on statistical data analysis and online analytical processing (OLAP). Descriptive analytics provides hindsight value regarding issues such as the identification of quantity and timing of current input flows [3], tracking of product location for accurate transportation [4,5], reverse logistics planning [6,7], and monitoring of emissions and air quality [8]. Predictive analytics uses identified trends to forecast maintenance [9], remanufacturing [10,11,12], process quality (e.g., through predicting resource needs, potential bottlenecks, and impact of interventions) [13], etc. Prescriptive analytics is a less commonly used area of business analytics [2]. Lieder et al. [14] proposed the use of machine learning (support vector machine classifiers) and simulation for assessing market acceptance in a circular economy context. Charnley et al. [15] developed the concept of certainty of product quality using discrete event simulation to depict the decision-making process of remanufacturing at the shop-floor level. In turn, Gbededo et al. [16] proposed the general theoretical framework for simulation-based sustainability impact analysis, however, without any consideration related to the aspect of its scalability and applicability. The scarcity of research devoted to a holistic approach that would merge predictive and prescriptive analytics was the motivation to develop an integrated approach dedicated to the improvement of sustainable manufacturing through supporting R&D specialists in identifying possible changes to be incorporated into product design. Moreover, the proposed approach should facilitate simulation performance and answer the question about prerequisites, by which a target level of sustainability performance will be achieved.
The aim of this study is to elaborate a method dedicated to seeking possibilities for changing product features with a goal of increasing sustainability of a product and its manufacturing. This method includes assessing sustainability performance, identifying relationships between factors that impact product sustainability, and, finally, performing simulations of changes in product features; these changes could increase sustainability performance. The proposed method is based on parametric models for identifying relationships and prescriptive analytics for using these relationships toward simulating sustainability performance. A simulation model can provide vast search space, which affects the time needed to obtain all possible solutions. To reduce processing time, a constraint programming technique is proposed that can be implemented within a constraint satisfaction problem (CSP). The advantage of the proposed method is the identification of all possible changes in the design process (if there are any), changes that can lead to the achievement of a target level of sustainability performance. Furthermore, constraint programming solves a problem according to the declarative paradigm facilitating the development of model specification.
Constraint satisfaction modeling is effectively used, for example, in conceptual design [17] and product cost evaluation [18,19]. However, this type of modeling is so far very rarely used in the field of assessing and improving sustainability performance. The application of CSP appears mainly in the aspect of managing sustainable supply chains [20,21]. Its utility in increasing sustainability of business through changes introduced during the design process has not yet been considered in the literature. This was an incentive to develop a method to support the R&D department in providing information regarding suitable changes in the design process, changes that could increase sustainability performance. The novelty of this study is twofold, related to theoretical and practical implications: firstly, the use of constraint satisfaction modeling to specify a model for the considered problem, and secondly, the development of a method for assessing sustainability performance and searching for possibilities of changing the design process toward improving sustainability performance in manufacturing. Consequently, the proposed method allows R&D specialists to obtain information about all possible changes (if there are any) at the design stage.
The framework of this chapter is as follows: Section 2 refers to a literature review including sustainable manufacturing, data-driven approaches, and business analytics for sustainable development. The proposed method of assessing sustainability performance and searching for possible changes in the design process is presented in Section 3. An illustrative example of using the proposed method is presented in Section 4. Finally, a conclusion and further research are described in Section 5.

2. Literature Review

2.1. Sustainable Manufacturing

The concept of sustainability refers to finding the trade-off between reaching environmental cleanliness, economic success, and social responsibility. These three aspects (environmental, economic, and social) are often called a triple-bottom-line framework. Sustainability from a company perspective is the meeting point between manufacturing, design, and environmental concerns. Nowadays, product sustainability tends to consider not only processes related to manufacturing and product usage, but also the product’s utilization, ultimate reuse, suitable recycling, and possible remanufacturing [22]. Closed-loop manufacturing systems aim to identify an efficient way to improve the flow of resources such as materials, components, and energy through multiple life cycles related to the single product [23].
Sustainable manufacturing is related to manufacturing products which minimize consumption of energy and natural resources (an environmental aspect), are profitable (an economic aspect), and are safe for customers, employees, and communities (a social aspect) [22]. When introducing sustainability to manufacturing companies, one should consider a holistic approach that includes issues related to the product, the manufacturing processes, the entire supply chain, and multiple life cycles of the product. Consequently, optimization techniques should be fitted to three levels regarding the product, process, and system. At the first level (the product level) the 6R concept (reduce, reuse, recycle, recover, redesign, and remanufacture) displaced the 3R concept (reduce, reuse, and recycle), enabling a change in the paradigm from the single life cycle to multiple life cycles. The second level (the process level) is related to achieving technological improvements by manufacturing companies to reduce consumption of resources (e.g., materials, energy, toxic wastes), increase product quality, and finally, extend the product life cycle. The third level (the system level) refers to all product life cycle phases (pre-manufacturing, manufacturing, usage, and post-usage), which are considered in the aspect of the entire supply chain and multiple life cycles of the product.
Design related to sustainability uses methods to improve design activities, materials, and supply chains toward creating a sustainable product. Sustainable products help businesses move toward sustainable manufacturing, in which the main aim is to manufacture high-quality products using more sustainable resources or fewer resources [24]. Sustainable manufacturing is possible if an enterprise takes into consideration the whole life cycle of the product, which extends beyond design, manufacturing, and product use [25]. Recently, sustainable manufacturing is often related to the concept of circular economy [26,27,28].
Product sustainability is often received by customers in the category of product appearance and eco-design, which can refer to a green color and eco-labelling. Issues such as product remanufacturing and product recovery are mostly neglected by customers [29]. Eco-design introduces environmental issues to a new product, enriching its other aspects regarding profit, image, functionality, quality, ergonomics, and aesthetics. The main aim of the eco-design concept is the reduction of resource consumption and negative environmental impacts, as well as an increase in product value [30].

2.2. Data-Driven Approaches and Business Analytics

Data-driven approaches are effectively used in many areas of sustainability, for example, in manufacturing [24,31,32], supply chain management [33], design [34,35,36], and e-mobility [37]. A framework of data analytics dedicated to sustainability performance can be seen from the perspective of four levels [31]: (1) data acquisition, (2) storage and preprocessing, (3) data mining, and (4) data application services. The first level refers to company databases, for instance, software related to computer-aided design and computer-aided engineering that supports the design process. The second level is related to retrieving data from company databases, data cleaning, integration, reduction, and transformation to a shape that can be used in data mining techniques. Data mining models are dedicated to a specific problem (for example, regression, classification, and clustering), and based on machine learning techniques using artificial neural networks, genetic algorithms, and support vector machines. The fourth level refers to supporting analysts in solving a considered problem, for instance, design optimization, energy optimization, and quality improvement. Data-driven approaches dedicated to the evaluation of sustainable manufacturing and product development are often based on methodologies referred to as an analytical hierarchy process [38,39,40,41] and a multi-criteria decision analysis [25,42,43,44,45,46].
The main aim of business analytics is to provide companies with significant insight and support management decisions regarding business performance, and consequently, gain a competitive advantage [47,48,49,50]. Moreover, business analytics can be seen as the extensive use of data, retrieved from various sources, statistical and quantitative analysis, as well as explanatory and predictive models, which are dedicated to providing information for decisions [1]. Business analytics is usually classified into three classes [49,51,52]: descriptive analytics, predictive analytics, and prescriptive analytics. Descriptive analytics attempts to answer to the question “What has happened?” Predictive analytics is related to answering the question “What might happen?” Finally, prescriptive analytics provides answers to questions such as “What should a business do to achieve its targets?” Descriptive analytics and predictive analytics are related to pattern identification in the past and future, respectively. The relationships between variables are mainly identified with the use of statistical analysis (e.g., linear regression and logistic regression) and machine learning (e.g., artificial neural networks and genetic algorithms). Prescriptive analytics suggests (prescribes) the best decision options, by which a company can take advantage of the predicted future with the use of large amounts of data. The outcome of prescriptive analytics can be considered within two aspects related to human intervention. The first aspect refers to decision support, for example, through providing recommendations. The second aspect is related to decision automation, for example, to implementing the prescribed actions. Prescriptive analytics tries to solve the abovementioned tasks using models and techniques belonging to operations research (including optimization algorithms, integer programming, dynamic programming, etc.), machine learning (including artificial neural networks, evolutionary computations, etc.), and simulations (including what-if scenarios, sensitivity analysis, etc.). These models and techniques aim to provide decision-makers with adaptive, automated, constrained, time-dependent, and above all optimal decisions in a probabilistic context [49,53]. Figure 1 illustrates descriptive, predictive, and prescriptive analytics in the context of their value for business and complexity.
Data-driven approaches should help the user in selecting and classifying evaluation indicators related to sustainability performance. These indicators should not only refer to the design optimization process (e.g., material properties, part dimensions, geometries), but also should include indicators regarding an environmental aspect (e.g., material consumption, energy consumption). In multi-criteria decision analysis, indicators are often divided into two groups [54,55,56,57]; the first group is related to the triple bottom line (environment, economy, society), and the second group refers to the process evaluation (e.g., material consumption, energy consumption, recyclability, repair rate, raw material cost, production cost, service cost).

3. A Method for Supporting Sustainable Manufacturing

The presented method aims to support R&D specialists in identifying possibilities to increase sustainability of manufacturing and a product through changes incorporated during the design process. This method provides information concerning opportunities for changes, for which sustainability performance of manufacturing may achieve the target level. The mentioned opportunities for changes can refer to reducing resource consumption and improving the manufacturing process, positively affecting environmental, economic, and societal aspects of sustainability.
The method consists of three steps: data collection, predictive analytics, and prescriptive analytics. Figure 2 illustrates a framework for the presented method.
As mentioned earlier, possibilities for improving sustainability performance refer to changes that could be introduced while developing a new product and that aim to increase product sustainability from the perspective of manufacturing and the product usage stage. The three steps of the proposed method are presented in detail below.

3.1. Data Collection

The first step of the proposed method refers to data collection mainly from company databases, to which access should be simplest. Company databases encompass many areas of business activities, including manufacturing, sales and marketing, design, accounting, warehouse management, and procurement. Manufacturing processes are supported by computer-integrated manufacturing that includes subsystems such as computer-aided design (CAD), computer-aided engineering (CAE), computer-aided manufacturing (CAM), computer-aided process planning (CAPP), computer-aided quality assurance (CAQ), production planning and control (PPC), and enterprise resource planning (ERP). These systems help employees manage the flow of raw materials, work-in-process inventory, finished goods, operation, machines, and information. Company databases can also register dates related to expectations obtained from potential consumers regarding to functionalities of a new product, and technological limitations recognized by R&D specialists.
Information on the amount of materials used in manufacturing, material quality, geometries, and dimensions of a product can be retrieved from CAD and CAE software. CAD enables designers to send some data (for example, bill of materials (BOM)) to ERP software in order to avoid time and mistakes made when inserting data related to BOM into an information system that supports manufacturing processes [58]. ERP comprises several modules that cover the majority of business activities, including manufacturing, sales, procurement, finance, warehouse management, human resources, and project management. The successful implementation of ERP enables a company to receive many benefits, for example, improved the flow of information, internal integration, and customer service. Many companies after an ERP system implementation have reported that their cost effectiveness in inventory, procurement, and productivity increased [59,60]. The module related to manufacturing includes data about processing times for specific parts of a product, and the number of finished products, including defective products.
The input variables are selected according to their significant impact on an output variable, which is the number of defective products. Moreover, input variables should be controllable in the context of easiness of changes, which can be carried out by top management to improve sustainability performance of the product and of manufacturing processes. An example of input variables can be the quality of materials used in manufacturing and the time of material processing.

3.2. Predictive Analytics

The second step of the proposed method is related to predictive analytics, in which data and variables are selected from company databases, relationships are identified, and predictions are made. The selection of variables and related data refers to a specific problem. Variables are divided into input (independent) variables and an output (dependent) variable, which is predicted, namely the number of defective products in manufacturing. The identification of relationships between input and output variables enables one to make predictions and simulations. The latter refers to the third step, namely prescriptive analytics.
This study uses parametric models based on regression analysis and artificial neural networks (ANNs). Linear or nonlinear modeling belonging to regression analysis allows the user to make predictions and inference of causal relationships. In turn, ANNs are part of nature-inspired methodologies called computational intelligence [18]. These methodologies are dedicated to solving complex real-world problems, for which the use of traditional models is often restricted. The advantages of regression analysis refer to implementation simplicity and fewer requirements for computational power compared to ANNs, particularly in the context of identifying nonlinear complex patterns.
In this study, both linear and nonlinear (polynomial) regression is used in the determination of predicted models. In turn, ANNs are trained according to gradient descent with momentum and adaptive learning rate backpropagation and Levenberg–Marquardt backpropagation. The most commonly used learning algorithm uses the backpropagation learning rule. This algorithm belongs to the supervised learning method. The dataset is divided into a training set and testing set to assess generalization capabilities of ANNs. These sets consist of input–output pairs regarding past data, which are processed by the learning algorithm. The quality of the abovementioned parametric models is assessed according to the mean absolute percentage error (MAPE):
M A P E % = 1 n ∑ t = 1 n x t − p t x t
where xt is the actual value, pt is the predicted value, and n is the number of input–output pairs (learning cases).
The verification of prediction quality requires the division of the dataset into the training and testing sets. Furthermore, experiments are performed using k-fold cross-validation, and the results are determined as the average of k-folds. Cross-validation is a resampling method, for which training and testing parametric models within k iterations is carried out according to the different subsets of data. Finally, predictions are made using the parametric model, for which the least MAPE in the testing set is achieved.

3.3. Prescriptive Analytics

In the third step of the proposed method, simulations are performed within specified variables, constraints, and according to an objective function to solve an optimization or decision-making problem. In this study, an objective is related to searching for opportunities for improving sustainability performance in manufacturing companies. The goal of simulations is to identify changes that could be introduced during the design process, aimed at increasing sustainability performance in manufacturing. In this study, the considered problem is formulated in terms of a CSP. The CSP facilitates development of a simulation environment through specifying three sets: variables {V1, V2, …, Vn}, discrete domains {D1, D2, …, Dn} related to variables V, and constraints {C1, C2, …, Cm}.
The CSP formalism can be described as a finite and discrete domain of values that is associated with each variable, whereas a constraint is a relationship that refers to a subset of the set of variables. The problem specification in terms of the CSP is an effective way to find a problem solution, if there are any. The solution of the CSP is a state, in which all constraints are satisfied. To achieve this state, learning mechanisms are performed through keeping valuable information for pruning the search space [61]. Constraint satisfaction modeling can be seen as a paradigm, which improves algorithmic techniques dedicated to solving real-life problems [62].
A set of solutions generated by the CSP reflects admissible changes in a new product, and it depends on the number of variables, their domains, and constraints, which interrelate variables. The constraints can refer to limitations within the technological process (e.g., the minimal density of materials), scarce resources (e.g., the number of machines), and relationships determined through parametric models. The problem specification in terms of the CSP allows R&D specialists to receive answers to questions regarding prediction and simulation:
  • What is the number of defective products predicted during the production process?
  • Is there a possibility for incorporating changes during the design process to increase sustainability performance, and if so, what changes are admissible?
In this study, the CSP is implemented using a constraint programming (CP) technique that involves constraint propagation and search algorithms. It is much easier to find a solution using a CSP after constraint propagation or to show that the CSP does not have a solution [63]. Constraint propagation can repeatedly reduce domains and/or constraints during its performance. Consequently, the CP reduces a CSP to an equivalent form that satisfies some local consistency notion [64]. The CP technique is particularly effective compared to an exhaustive search that finds a solution if one exists, but its performance is proportional to the number of admissible solutions. As a result, an exhaustive search tends to increase very quickly according to the size of a problem, limiting its application in many practical problems [62].

4. An Example of Applying the Proposed Method

An example illustrates the applicability of the proposed method, and it is divided into three subsections: model specification, predictive analytics, and prescriptive analytics. The presented example refers to reducing the number of defective products in manufacturing through increasing material quality for products. The improvement of material quality also affects the long-term and intensive usage of a product, increasing sustainability performance of products and manufacturing processes.

4.1. Model Specification

The reduction of the number of defective products in the manufacturing process and, consequently, the reduction of material consumption improves sustainability performance regarding the environmental and economic aspects of manufacturing sustainability. Variable selection has been based on the author’s experience and a literature analysis [15,65,66]. The model specification for the considered problem includes the following variables:
  • V1—the total time of material processing (in minutes);
  • V2—the size of components for material processing (in cm3);
  • V3—the material density (in g/cm3);
  • V4—the number of defective products;
  • V5—the unit cost related to defective products (in EUR);
  • V6—the unit production cost (in EUR);
  • V7—the material cost per product (in EUR);
  • V8—the labor cost per product (in EUR);
  • V9—the energy cost per product (in EUR);
  • V10—the overhead cost per product (in EUR).
The relationship between the number of defective products and input variables is determined as follows:
V4 = f(V1, V2, V3)
The unit production cost is calculated according to the following formula:
V6 = V7 + V8 + V9 + V10
The values for variables V1–V3 and V7–V10 are determined using an analogical approach, finding production parameters and cost performance of the previous product most similar to the new product. The identification of the most similar product can include criteria referring to product size, weight, features, and functionalities of the new product. Equation (2) is determined with the use of parametric modeling. Statistical significance shows that all input variables (V1–V3) are significant for the selected regression models by the 5% significance level.

4.2. Predictive Analytics

The sample size consists of 21 cases regarding past and existing products belonging to the same product line as a new product. To verify prediction quality of parametric models, the dataset was divided into two sets: learning and testing. The learning set includes records regarding 16 products, and the testing set includes 5 records. The 5-fold cross-validation was used during experiments. The number of defective products was determined using parametric models based on linear regression (LR), polynomial regression (PR), and artificial neural networks trained according to the gradient descent with momentum and adaptive learning rate backpropagation algorithm (NN-GD), and Levenberg–Marquardt backpropagation algorithm (NN-LM). The results obtained by parametric models were compared to the average (AV) of an output variable—the number of defective products. Table 1 presents the comparison of MAPEs for various prediction models for the learning and testing sets.
The results of experiments presented in Table 1 show that prediction models related to ANNs generate lower MAPEs in the testing set compared to regression analysis. All parametric models outperform prediction quality for the average of the number of defective products. The least MAPE in the testing set was obtained by the NN-GD model. Therefore, this model was used to predict the number of defective products in manufacturing. Assuming the following values of input variables: V1 = 94, V2 = 286, V3 = 7.3, the number of defective products was predicted at 26 units per 1000 manufactured products.

4.3. Prescriptive Analytics

Prescriptive analytics refers to the application of the constraint programming technique within the considered CSP. The advantage of using constraint programming is the significant reduction of the computational time compared to an exhaustive search. It is particularly important for the vast search space in order to find all admissible solutions. Simulations were carried out according to the CSP environment, in which the variables, their domains, and constraints are specified. Constraints are linked with accessible resources and Equations (2) and (3). The goal of simulations is to find opportunities (if there are any) to reduce the number of defective products through suitable changes in the total time of material processing and material density. Equation (2) indicates that the increase in material density improves product quality, and as a result, the number of defective products is reduced. Consequently, fewer defective products in manufacturing reduce material consumption, labor, energy consumption, and related costs.
Table 2 presents results of a few simulations that refer to the reduction of the number of defective products and the unit cost related to defective products, material, and production. Simulations were carried out for two variables that can be controlled by the company in the manufacturing process: the total time of material processing (V1) and material density (V3). For these variables, the following domains were assumed: D1 = {85, …, 94} and D3 = {7.3, …, 8.2}. As a result, there were 100 potential changes in the manufacturing process by which the number of defective products could be reduced. Simulations were performed using the Mozart programming system that includes constraint programming techniques.
The results presented in Table 2 show that the increase in material density from 7.3 to 8.2 g/cm3 reduces the number of defective products by about 4 products (from 26 to 22 products for the total time of material processing at 94 min). As a result, the cost of manufacturing defective products is reduced, including material consumption, labor, and energy consumption related to the production processes. On the other hand, the unit production cost increases after improving material density, resulting in the greater material cost. However, the smaller number of defective products affects not only the manufacturing phase, but also the product usage phase, reducing the warranty cost and increasing customer satisfaction and loyalty. In this case, the support dedicated to R&D specialists for finding suitable changes to reduce the number of defective products seems to be particularly useful. The presented simulations can also be expanded toward obtaining the answer to a question about prerequisites, for which sustainability performance reaches the highest level. For instance, the minimal number of defective products (i.e., 17 products) is determined at material density of 8.2 g/cm3 and the total time of material processing of 85 min. Moreover, further analysis could be expanded toward searching for prerequisites, for which the desirable level of the cost could be achieved.

5. Discussion

The present study is concerned with developing an approach to merging predictive and prescriptive analytics dedicated to improving sustainable manufacturing. Predictive analytics is based on parametric modeling using regression analysis and artificial neural networks. The results of experiments show that ANNs are able to identify nonlinear relationships among data more precisely than regression models. This is similar to research carried out in other areas related to sustainable manufacturing such as modeling of turning operations [67], fault diagnosis [68], and maintenance planning [69]. Gue et al. [70] indicate that ANNs are used as an advanced tool for modeling complex behavior of systems. The results obtained in the present study are consistent with this statement, using ANNs to identifying trends and predicting sustainability performance of a product and manufacturing system.
Prescriptive analytics, including issues related to simulations of sustainability performance in manufacturing, is a less commonly investigated field of research [2]. There have been reported some studies regarding the use of machine learning to simulate remanufacturing [71,72], energy consumption [73,74], and in the context of Industry 4.0 [24,75]. Gbededo et al. [16] proposed the general theoretical framework for simulation-based sustainability impact analysis. However, these studies have not included the merging of predictive and prescriptive analytics into a holistic approach and the applicability of this approach in improving sustainable manufacturing. The present study considers the use of constraint satisfaction modeling to build a simulation environment of searching for possibilities of changes in the design process to increase sustainability performance of a product and manufacturing. The advantage of using constraint satisfaction modeling is the declarative nature of constraint programming, and consequently, the easiness of developing a model for the considered problem through adding and/or removing variables and constraints. Constraint satisfaction modeling can also be seen as a pertinent framework for developing a decision support system. Moreover, theoretical implications can be considered in the context of finding answers to two types of questions: about the value of an output variable (sustainability performance), and about prerequisites (the values of input variables) by which a target level of an output variable can be reached.
The main contribution of the present research is related to using a CSP paradigm to the specification of the considered problem toward obtaining all admissible improvement opportunities (if there are any) of sustainability performance. The problem formulation in terms of a CSP enables the application of a constraint programming, and consequently, the time-effective reduction of the search space, which ensures interactive properties of a decision support system dedicated to the considered problem. Moreover, the problem formulation in terms of a CSP facilitates the gradual extension of a knowledge base that consists of facts, constraints, and relationships described in the form of if–then rules. Furthermore, the use of a CSP paradigm enables effortless development of the model toward describing another optimization or decision problem, for instance, an energy consumption problem.
The practical implications of the present research are found in the proposed method of assessing sustainability performance and searching for opportunities to implement changes in the design process with a goal of improving sustainability performance in manufacturing. The proposed method allows R&D specialists to obtain information about all possible changes (if there are any) at the product design stage that proceeds the manufacturing process and later stages of the product life cycle. As a result, incorporating changes in a product during its design enables cost reduction in the successive stages. For example, the cost of defective products considered in this study can affect the manufacturing process and after-sales service, and consequently, affect the direct production cost and warranty cost. Finally, the managerial implications include the support for the R&D department in searching for improvement opportunities of product and manufacturing sustainability, and the support for top management in increasing manufacturing effectiveness through reducing the number of defective products.

6. Conclusions

The improvement of sustainability performance considered in this study refers to the increase in material quality, and consequently, the reduction of the number of defective products in manufacturing. This improvement refers also to the reduction of materials, labor, and energy at the manufacturing stage. The proposed method evaluates sustainability performance and supports R&D specialists in searching for improvement opportunities in the design process, for which sustainability performance of a new product and manufacturing could be increased. Moreover, constraint satisfaction modeling enables the identification of all possible changes in the design and manufacturing process to improve sustainability performance. These changes can be something new and interesting for R&D specialists, and can help them delineate new directions for modifications of the product and manufacturing process. However, a simulation environment can produce an enormous number of solutions that may exceed human abilities to interpret the simulation results. Another limitation of the proposed approach refers to the issue of company databases that store specifications related to past and existing products. Hence, the proposed method is not intended for supporting the design process related to innovations. The process of collecting data only from company databases and data preprocessing for use in machine learning techniques can also be considered as a disadvantage of the proposed approach. In addition to this, company databases usually include the small number of similar products within a single product line. As a result, the statistical analysis related to identifying relationships may be hard to interpret.
Future research can partially address the above limitations. Firstly, an enormous number of solutions can be reduced through the increase in granularity related to decision variables. Further research can also develop a decision support system to help the user select the most suitable parametric modeling method to identify patterns, and view results according to the user’s preferences, for instance, the maximum sustainability performance of a product and manufacturing. Another direction of future research could address the cost analysis related to reducing the number of defective products in the context of increasing prices of materials. This direction can be expanded toward receiving information from potential customers about their preferences regarding product features in the context of the product’s sustainability, and their purchasing preferences, including the maximum price of a new, more sustainable product.

Funding

This research received no external funding.

Institutional Review Board Statement

Not applicable.

Informed Consent Statement

Not applicable.

Data Availability Statement

Not applicable.

Conflicts of Interest

The author declares no conflict of interest.

References

  1. Davenport, T.; Harris, J. Competing on Analytics: Updated, with a New Introduction: The New Science of Winning; Harvard Business Press: Brighton, MA, USA, 2017. [Google Scholar]
  2. Kristoffersen, E.; Blomsma, F.; Mikalef, P.; Li, J. The smart circular economy: A digital-enabled circular strategies framework for manufacturing companies. J. Bus. Res. 2020, 120, 241–261. [Google Scholar] [CrossRef]
  3. Pagoropoulos, A.; Pigosso, D.C.; McAloone, T.C. The emergent role of digital technologies in the circular economy: A review. Procedia CIRP 2017, 64, 19–24. [Google Scholar] [CrossRef]
  4. Porter, M.E.; Heppelmann, J.E. How smart, connected products are transforming competition. Harv. Bus. Rev. 2014, 92, 64–88. [Google Scholar]
  5. Romero, D.; Noran, O. Towards green sensing virtual enterprises: Interconnected sensing enterprises, intelligent assets and smart products in the cyber-physical circular economy. IFAC PapersOnLine 2017, 50, 11719–11724. [Google Scholar] [CrossRef]
  6. Govindan, K.; Soleimani, H.; Kannan, D. Reverse logistics and closed-loop supply chain: A comprehensive review to explore the future. Eur. J. Oper. Res. 2015, 240, 603–626. [Google Scholar] [CrossRef]
  7. Zhou, Z.; Cai, Y.; Xiao, Y.; Chen, X.; Zeng, H. The optimization of reverse logistics cost based on value flow analysis–a case study on automobile recycling company in China. J. Intell. Fuzzy Syst. 2018, 34, 807–818. [Google Scholar] [CrossRef]
  8. Molka-Danielsen, J.; Engelseth, P.; Wang, H. Large scale integration of wireless sensor network technologies for air quality monitoring at a logistics shipping base. J. Ind. Inf. Integr. 2018, 10, 20–28. [Google Scholar] [CrossRef]
  9. Li, Z.; Wang, Y.; Wang, K.S. Intelligent predictive maintenance for fault diagnosis and prognosis in machine centers: Industry 4.0 scenario. Adv. Manuf. 2017, 5, 377–387. [Google Scholar] [CrossRef]
  10. Liu, C.; Zhu, Q.; Wei, F.; Rao, W.; Liu, J.; Hu, J.; Cai, W. A review on remanufacturing assembly management and technology. Int. J. Adv. Manuf. Technol. 2019, 105, 4797–4808. [Google Scholar] [CrossRef]
  11. Bressanelli, G.; Adrodegari, F.; Perona, M.; Saccani, N. Exploring how usage-focused business models enable circular economy through digital technologies. Sustainability 2018, 10, 639. [Google Scholar] [CrossRef]
  12. Yang, S.; Araving Raghavendra, M.R.; Kaminski, J.; Pepin, H. Opportunities for industry 4.0 to support remanufacturing. Appl. Sci. 2018, 8, 1177. [Google Scholar] [CrossRef]
  13. Lin, K.P.; Yu, C.M.; Chen, K.S. Production data analysis system using novel process capability indices-based circular economy. Ind. Manag. Data Syst. 2019, 119, 1655–1668. [Google Scholar] [CrossRef]
  14. Lieder, M.; Asif, F.M.; Rashid, A. A choice behavior experiment with circular business models using machine learning and simulation modeling. J. Clean. Prod. 2020, 258, 120894. [Google Scholar] [CrossRef]
  15. Charnley, F.; Tiwari, D.; Hutabarat, W.; Moreno, M.; Okorie, O.; Tiwari, A. Simulation to enable a data-driven circular economy. Sustainability 2019, 11, 3379. [Google Scholar] [CrossRef]
  16. Gbededo, M.A.; Liyanage, K.; Garza-Reyes, J.A. Towards a life cycle sustainability analysis: A systematic review of approaches to sustainable manufacturing. J. Clean. Prod. 2018, 184, 1002–1015. [Google Scholar] [CrossRef]
  17. Li, M.; Lou, S.; Gao, Y.; Zheng, H.; Hu, B.; Tan, J. A cerebellar operant conditioning-inspired constraint satisfaction approach for product design concept generation. Int. J. Prod. Res. 2022, 1–20. [Google Scholar] [CrossRef]
  18. Relich, M. Computational intelligence for estimating cost of new product development. Found. Manag. 2016, 8, 21–34. [Google Scholar] [CrossRef]
  19. Wang, J.; Li, R.; Ding, G.; Qin, S.; Cai, Z. Product-service system engineering characteristics design for life cycle cost based on constraint satisfaction problem and Bayesian network. Adv. Eng. Inform. 2022, 52, 101573. [Google Scholar] [CrossRef]
  20. Lahri, V.; Shaw, K.; Ishizaka, A. Sustainable supply chain network design problem: Using the integrated BWM, TOPSIS, possibilistic programming, and ε-constrained methods. Expert Syst. Appl. 2021, 168, 114373. [Google Scholar] [CrossRef]
  21. Shahparvari, S.; Soleimani, H.; Govindan, K.; Bodaghi, B.; Fard, M.T.; Jafari, H. Closing the loop: Redesigning sustainable reverse logistics network in uncertain supply chains. Comput. Ind. Eng. 2021, 157, 107093. [Google Scholar] [CrossRef]
  22. Sartal, A.; Bellas, R.; Mejías, A.M.; García-Collado, A. The sustainable manufacturing concept, evolution and opportunities within Industry 4.0: A literature review. Adv. Mech. Eng. 2020, 12, 1687814020925232. [Google Scholar] [CrossRef]
  23. Andersen, A.L.; Brunoe, T.D.; Bockholt, M.T.; Napoleone, A.; Kristensen, J.H.; Colli, M.; Nielsen, K. Changeable closed-loop manufacturing systems: Challenges in product take-back and evaluation of reconfigurable solutions. Int. J. Prod. Res. 2023, 61, 839–858. [Google Scholar] [CrossRef]
  24. Machado, C.G.; Winroth, M.P.; Ribeiro da Silva, E.H. Sustainable manufacturing in Industry 4.0: An emerging research agenda. Int. J. Prod. Res. 2020, 58, 1462–1484. [Google Scholar] [CrossRef]
  25. Zarte, M.; Pechmann, A.; Nunes, I.L. Decision support systems for sustainable manufacturing surrounding the product and production life cycle: A literature review. J. Clean. Prod. 2019, 219, 336–349. [Google Scholar] [CrossRef]
  26. Hapuwatte, B.M.; Jawahir, I.S. Closed-loop sustainable product design for circular economy. J. Ind. Ecol. 2021, 25, 1430–1446. [Google Scholar] [CrossRef]
  27. Agrawal, R.; Wankhede, V.A.; Kumar, A.; Upadhyay, A.; Garza-Reyes, J.A. Nexus of circular economy and sustainable business performance in the era of digitalization. Int. J. Product. Perform. Manag. 2022, 71, 748–774. [Google Scholar] [CrossRef]
  28. Bag, S.; Pretorius, J.H. Relationships between industry 4.0, sustainable manufacturing and circular economy: Proposal of a research framework. Int. J. Organ. Anal. 2022, 30, 864–898. [Google Scholar] [CrossRef]
  29. Diego-Mas, J.A.; Poveda-Bautista, R.; Alcaide-Marzal, J. Designing the appearance of environmentally sustainable products. J. Clean. Prod. 2016, 135, 784–793. [Google Scholar] [CrossRef]
  30. Rau, H.; Lagapa, M.D.; Chen, P.H. Anticipatory non-green-phenomena determination for designing eco-design products. Sustainability 2021, 13, 621. [Google Scholar] [CrossRef]
  31. Majeed, A.; Zhang, Y.; Ren, S.; Lv, J.; Peng, T.; Waqar, S.; Yin, E. A big data-driven framework for sustainable and smart additive manufacturing. Robot. Comput. Integr. Manuf. 2021, 67, 102026. [Google Scholar] [CrossRef]
  32. Ma, S.; Ding, W.; Liu, Y.; Ren, S.; Yang, H. Digital twin and big data-driven sustainable smart manufacturing based on information management systems for energy-intensive industries. Appl. Energy 2022, 326, 119986. [Google Scholar] [CrossRef]
  33. Tsai, F.M.; Bui, T.D.; Tseng, M.L.; Ali, M.H.; Lim, M.K.; Chiu, A.S. Sustainable supply chain management trends in world regions: A data-driven analysis. Resour. Conserv. Recycl. 2021, 167, 105421. [Google Scholar] [CrossRef]
  34. Kishawy, H.A.; Hegab, H.; Saad, E. Design for sustainable manufacturing: Approach, implementation, and assessment. Sustainability 2018, 10, 3604. [Google Scholar] [CrossRef]
  35. Furxhi, I.; Perucca, M.; Blosi, M.; Lopez de Ipiña, J.; Oliveira, J.; Murphy, F.; Costa, A.L. ASINA Project: Towards a Methodological Data-Driven Sustainable and Safe-by-Design Approach for the Development of Nanomaterials. Front. Bioeng. Biotechnol. 2022, 9, 1456. [Google Scholar] [CrossRef]
  36. Sun, H.; Guo, W.; Wang, L.; Lin, M. A data-driven adaptive design for achieving sustainable product. Procedia CIRP 2022, 109, 316–321. [Google Scholar] [CrossRef]
  37. Nigro, M.; Ferrara, M.; De Vincentis, R.; Liberto, C.; Valenti, G. Data driven approaches for sustainable development of E-mobility in urban areas. Energies 2021, 14, 3949. [Google Scholar] [CrossRef]
  38. Ocampo, L.A.; Vergara, V.G.; Impas, C.; Tordillo, J.A.; Pastoril, J. Identifying critical indicators in sustainable manufacturing using analytic hierarchy process (AHP). J. Manuf. Ind. Eng. 2015, 14, 1–8. [Google Scholar] [CrossRef]
  39. Askary, Z.; Singh, A.; Gupta, S.; Shukla, R.K.; Jaiswal, P. Development of AHP Framework of Sustainable Product Design and Manufacturing of Electric Vehicle. In Advances in Engineering Design. Lecture Notes in Mechanical Engineering; Prasad, A., Gupta, S., Tyagi, R., Eds.; Springer: Singapore, 2019; pp. 415–422. [Google Scholar]
  40. Deshmukh, R.A.; Hiremath, R. Analyzing the Key Performance Indicators of Advanced Sustainable Manufacturing System Using AHP Approach. In Techno-Societal 2018; Pawar, P., Ronge, B., Balasubramaniam, R., Vibhute, A., Apte, S., Eds.; Springer: Cham, Switzerland, 2020; pp. 745–750. [Google Scholar]
  41. Saraf, C.; Agrawal, S.; Barodiya, D.; Shrivastava, P.; Verma, T.N. AHP-based Identification of Tools for Sustainable Product Development. In Advancement in Materials, Manufacturing and Energy Engineering; Springer Nature: Singapore, 2022; pp. 543–552. [Google Scholar]
  42. Kluczek, A. An overall multi-criteria approach to sustainability assessment of manufacturing processes. Procedia Manuf. 2017, 8, 136–143. [Google Scholar] [CrossRef]
  43. Mathiyazhagan, K.; Gnanavelbabu, A.; Lokesh Prabhuraj, B. A sustainable assessment model for material selection in construction industries perspective using hybrid MCDM approaches. J. Adv. Manag. Res. 2019, 16, 234–259. [Google Scholar] [CrossRef]
  44. Bhanot, N.; Qaiser, F.H.; Alkahtani, M.; Rehman, A.U. An integrated decision-making approach for cause-and-effect analysis of sustainable manufacturing indicators. Sustainability 2020, 12, 1517. [Google Scholar] [CrossRef]
  45. Hichem, A.; Mohyeddine, S.; Abdessamed, K. Benchmarking framework for sustainable manufacturing based MCDM techniques. Benchmarking Int. J. 2022, 29, 87–117. [Google Scholar] [CrossRef]
  46. Guebsi, W.; Zouari, A. Investigation on multi-criteria decision making methods application in sustainable product design. SPEKTA 2022, 3, 91–104. [Google Scholar] [CrossRef]
  47. Vidgen, R.; Shaw, S.; Grant, D.B. Management challenges in creating value from business analytics. Eur. J. Oper. Res. 2017, 261, 626–639. [Google Scholar] [CrossRef]
  48. Singh, S.K.; El-Kassar, A.N. Role of big data analytics in developing sustainable capabilities. J. Clean. Prod. 2019, 213, 1264–1273. [Google Scholar] [CrossRef]
  49. Lepenioti, K.; Bousdekis, A.; Apostolou, D.; Mentzas, G. Prescriptive analytics: Literature review and research challenges. Int. J. Inf. Manag. 2020, 50, 57–70. [Google Scholar] [CrossRef]
  50. Relich, M.; Bzdyra, K. Knowledge discovery in enterprise databases for forecasting new product success. Lect. Notes Comput. Sci. 2015, 9375, 121–129. [Google Scholar]
  51. Krumeich, J.; Werth, D.; Loos, P. Prescriptive control of business processes. Bus. Inf. Syst. Eng. 2016, 58, 261–280. [Google Scholar] [CrossRef]
  52. Sun, K.; Li, Y.; Roy, U. A PLM-based data analytics approach for improving product development lead time in an engineer-to-order manufacturing firm. Math. Model. Eng. Probl. 2017, 4, 69–74. [Google Scholar] [CrossRef]
  53. Bertsimas, D.; Kallus, N. From predictive to prescriptive analytics. Manag. Sci. 2020, 66, 1025–1044. [Google Scholar] [CrossRef]
  54. Jiang, Z.; Ding, Z.; Zhang, H.; Cai, W.; Liu, Y. Data-driven ecological performance evaluation for remanufacturing process. Energy Convers. Manag. 2019, 198, 111844. [Google Scholar] [CrossRef]
  55. Li, T.; Yeo, J. Strengthening the sustainability of additive manufacturing through data-driven approaches and workforce development. Adv. Intell. Syst. 2021, 3, 2100069. [Google Scholar] [CrossRef]
  56. Perdeli Demirkan, C.; Smith, N.M.; Duzgun, H.S.; Waclawski, A. A data-driven approach to evaluation of sustainability reporting practices in extractive industries. Sustainability 2021, 13, 8716. [Google Scholar] [CrossRef]
  57. Wang, L.; Liu, Z. Data-driven product design evaluation method based on multi-stage artificial neural network. Appl. Soft Comput. 2021, 103, 107117. [Google Scholar] [CrossRef]
  58. Li, M.; Li, L.; Ma, Y. Integration of well-defined BIM external module with CAD via associative feature templates. Comput. Aided Des. Appl. 2019, 16, 878–893. [Google Scholar] [CrossRef]
  59. Dezdar, S. An integrative model for realising benefits from enterprise resource planning implementation. Int. J. Bus. Inf. Syst. 2017, 24, 423–451. [Google Scholar] [CrossRef]
  60. Menon, S. Benefits and process improvements for ERP implementation: Results from an exploratory case study. Int. Bus. Res. 2019, 12, 124–132. [Google Scholar] [CrossRef]
  61. Ghedira, K. Constraint Satisfaction Problems: CSP Formalisms and Techniques; John Wiley & Sons: Hoboken, NJ, USA, 2013. [Google Scholar]
  62. Relich, M.; Świć, A. Parametric estimation and constraint programming-based planning and simulation of production cost of a new product. Appl. Sci. 2020, 10, 6330. [Google Scholar] [CrossRef]
  63. Rossi, F.; Van Beek, P.; Walsh, T. Constraint programming. Found. Artif. Intell. 2008, 3, 181–211. [Google Scholar]
  64. Relich, M. Identifying project alternatives with the use of constraint programming. Adv. Intell. Syst. Comput. 2016, 521, 3–13. [Google Scholar]
  65. Hassan, M.F.; Saman, M.Z.; Sharif, S.; Omar, B. Integration of morphological analysis theory and artificial neural network approach for sustainable product design: A case study of portable vacuum cleaner. Int. J. Sustain. Manuf. 2012, 2, 293–316. [Google Scholar] [CrossRef]
  66. Lahmar, H.; Dahane, M.; Mouss, N.K.; Haoues, M. Production planning optimisation in a sustainable hybrid manufacturing remanufacturing production system. Procedia Comput. Sci. 2022, 200, 1244–1253. [Google Scholar] [CrossRef]
  67. Cica, D.; Sredanovic, B.; Tesic, S.; Kramar, D. Predictive modeling of turning operations under different cooling/lubricating conditions for sustainable manufacturing with machine learning techniques. Appl. Comput. Inform. 2020. ahead of print. [Google Scholar] [CrossRef]
  68. Grezmak, J.; Zhang, J.; Wang, P.; Gao, R.X. Multi-stream convolutional neural network-based fault diagnosis for variable frequency drives in sustainable manufacturing systems. Procedia Manuf. 2020, 43, 511–518. [Google Scholar] [CrossRef]
  69. Abidi, M.H.; Mohammed, M.K.; Alkhalefah, H. Predictive maintenance planning for industry 4.0 using machine learning for sustainable manufacturing. Sustainability 2022, 14, 3387. [Google Scholar] [CrossRef]
  70. Gue, I.H.; Ubando, A.T.; Tseng, M.L.; Tan, R.R. Artificial neural networks for sustainable development: A critical review. Clean Technol. Environ. Policy 2020, 22, 1449–1465. [Google Scholar] [CrossRef]
  71. Goodall, P.; Sharpe, R.; West, A. A data-driven simulation to support remanufacturing operations. Comput. Ind. 2019, 105, 48–60. [Google Scholar] [CrossRef]
  72. He, P. Optimization and simulation of remanufacturing production scheduling under uncertainties. Int. J. Simul. Model. 2018, 17, 734–743. [Google Scholar] [CrossRef]
  73. Garwood, T.L.; Hughes, B.R.; Oates, M.R.; O’Connor, D.; Hughes, R. A review of energy simulation tools for the manufacturing sector. Renew. Sustain. Energy Rev. 2018, 81, 895–911. [Google Scholar] [CrossRef]
  74. Relich, M.; Gola, A.; Jasiulewicz-Kaczmarek, M. Identifying improvement opportunities in product design for reducing energy consumption. Energies 2022, 15, 9611. [Google Scholar] [CrossRef]
  75. Enyoghasi, C.; Badurdeen, F. Industry 4.0 for sustainable manufacturing: Opportunities at the product, process, and system levels. Resour. Conserv. Recycl. 2021, 166, 105362. [Google Scholar] [CrossRef]
Sustainability 15 07667 g001 550
Figure 1. Types of business analytics.
Figure 1. Types of business analytics.
Sustainability 15 07667 g001
Sustainability 15 07667 g002 550
Figure 2. A framework for the presented method.
Figure 2. A framework for the presented method.
Sustainability 15 07667 g002
Table
Table 1. The comparison of MAPEs for prediction models (in %).
Table 1. The comparison of MAPEs for prediction models (in %).
Prediction ModelLearning SetTesting Set
LR13.1114.97
PR11.8412.73
NN-GD12.0410.83
NN-LM6.5912.31
AV19.7123.08
Table
Table 2. Selected results of simulations.
Table 2. Selected results of simulations.
VariablesV4V5V7V6
V1 = 85, V3 = 7.3 2163.1148.1246.9
……………
V1 = 94, V3 = 7.32673.8148.9254.8
……………
V1 = 85, V3 = 7.42062.5150.2248.8
……………
V1 = 94, V3 = 7.42573.2151.0256.7
……………
V1 = 85, V3 = 8.21755.6166.8263.6
……………
V1 = 94, V3 = 8.22265.8172.0271.5
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Relich, M. Predictive and Prescriptive Analytics in Identifying Opportunities for Improving Sustainable Manufacturing. Sustainability 2023, 15, 7667. https://doi.org/10.3390/su15097667

AMA Style

Relich M. Predictive and Prescriptive Analytics in Identifying Opportunities for Improving Sustainable Manufacturing. Sustainability. 2023; 15(9):7667. https://doi.org/10.3390/su15097667

Chicago/Turabian Style

Relich, Marcin. 2023. "Predictive and Prescriptive Analytics in Identifying Opportunities for Improving Sustainable Manufacturing" Sustainability 15, no. 9: 7667. https://doi.org/10.3390/su15097667

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop