Skip to main content
Top

2024 | Book

Analytical Decision Making and Data Envelopment Analysis

Advances and Challenges

Editors: S.A Edalatpanah, Farhad Hosseinzadeh Lotfi, Kristiaan Kerstens, Peter Wanke

Publisher: Springer Nature Singapore

Book Series : Infosys Science Foundation Series

insite
SEARCH

About this book

This book explores the intersection of data envelopment analysis (DEA) and various analytical decision-making methodologies. Featuring contributions from experts in the field from across the world, each chapter delves into different aspects of DEA and its applications in real-world scenarios. The book covers a wide range of topics, including integrating DEA with machine learning techniques, performance evaluation in diverse sectors like banking and civil engineering, and using DEA in managerial decision-making. It also examines data mining during the Covid-19 pandemic and the application of blockchain and IoT in supply chain management. The book offers a deep dive into the evolution of nonparametric frontier methods and the development of new optimization algorithms, addressing the complexities of modern analytical decision-making tools.

A few chapters delve into futuristic topics like fuzzy sets and their extensions in decision-making and exploring e-learning platforms for education. This book is an invaluable resource for researchers, practitioners and students interested in the latest DEA advancements and practical applications in various fields. Its multidisciplinary approach makes it a useful addition to the libraries of those seeking to understand the complexities and potentials of modern analytical decision-making tools.

Table of Contents

Frontmatter
Merging Data Envelopment Analysis and Structural Risk Minimization: Some Examples of Use of Multi-output Machine Learning Techniques on Real-World Data
Abstract
Data Envelopment Analysis (DEA) is nowadays a very famous nonparametric technique for the measurement of technical efficiency. It does so by building a production possibility set that satisfies certain microeconomic and mathematical axioms, such as free disposability in inputs and outputs and convexity, and determines the most conservative estimate of technical inefficiency of each assessed unit via the minimal extrapolation principle (i.e., the Occam's Razor view). Given a data sample and from a statistical point of view, this last axiom implies an estimation of technical inefficiency by exclusively minimizing the empirical error, resulting in overfitting to the data as a by-product. To overcome this statistical problem when the objective is measuring technical inefficiency beyond the data sample, a methodology has been recently introduced which follows the Structural Risk Minimization (SRM) principle. It controls both the empirical and the generalization (prediction) error of the model. This methodology, called Data Envelopment Analysis-based Machines (DEAM) was introduced for the single-output setting [(Guerrero, Aparicio, & Valero-Carreras 2022). Combining Data Envelopment Analysis and Machine Learning. Mathematics 2022, 10, 909.]. In this chapter, we extend the DEAM approach for the estimation of production functions to the multi-output production framework, evaluate technical efficiency with respect to a variety of measures, and illustrate its performance via some empirical applications. In this way, we provide some examples of use of multi-output machine learning techniques for measuring technical efficiency on real-world data.
Nadia M. Guerrero, Juan Aparicio, Raul Moragues, Daniel Valero-Carreras
A New Network Data Envelopment Analysis Model for Efficacy Evaluation of Decision-Making Units
Abstract
Assessing the performance of organizations has become a complex matter within the realm of management, considering the current global setting and its ever-changing nature. Although Data Envelopment Analysis (DEA) has been used as a powerful tool for the evaluation of decision-making units, in the traditional DEA models, these units are viewed as a black box, and therefore their internal relationships are generally ignored, assuming that the performance of DMUs is a function of selective inputs and outputs. Recently, there has been a great deal of research on multi-stage units, especially series and parallel units. However, research with a mixed structure is very limited with many studies unable to introduce the efficient unit. However, since in most real cases, the researcher deals with a series and parallel structure simultaneously, it is vital to develop a model of efficiency evaluation with efficient units. Thus, in the current paper, we attempted to solve the mentioned problems. At first, a mixed structure was introduced and then we developed a new model for a mixed structure. Also, the proposed model has the capability of introducing a relative-efficient unit. This is followed by an explanation of the methodology and development of a new network model to measure efficiency in the Iranian Airline Industry.
Hilda Saleh, Morteza Shafiee, Sara Nakhjirkan
Possibilistic Network DEA Approach for Performance Evaluation of Two-Stage Decision Making Units Under Uncertainty
Abstract
Uncertainty is a significant context to investigate when assessing entities. In the presence of imprecise and vague data, this study presents a novel method for evaluating the performance of decision-making units (DMUs) utilizing a network structure consisting of two stages. To present the fuzzy network data envelopment analysis (FNDEA) model, two-stage data envelopment analysis (TSDEA), chance-constrained programming (CCP), and possibilistic programming are employed. Furthermore, the possibilistic network data envelopment analysis (PNDEA) method  can be utilized under various returns to scale (RTS) assumptions. To measure the performance of investment firms (IFs) using a two-stage structure that includes portfolio and operational management procedures, the developed fuzzy network DEA model is implemented. In addition, IFs like mutual funds (MFs) and investment organizations are extremely significant organizations to make investments in capital markets. Consequently, assessing the performance of these firms to identify efficient IFs and propose appropriate solutions for inefficient IFs is important. Finally, a real-world case study from the Tehran Stock Exchange is used, and the findings indicate that the developed model is effective.
Pejman Peykani, Mostafa Sargolzaei, Farhad Hamidzadeh, Fatemeh Sadat Seyed Esmaeili, Amir Takaloo
Managerial Ability in Indian Life Insurance Companies: A Comparison Based on DEA and DEAGP
Abstract
Although global life insurance industry experienced significant structural and technological changes over the past decade, it is expected to increase based on the next decade’s outlook. The performance of life insurance companies is influenced by several factors, and managerial ability is a key facilitator for the companies aiming to gain a leadership position in insurance market. Thus managerial ability evaluation is essential for retaining and rewarding capable managers. Nevertheless, managerial ability is an intellectual capability that could not be measured directly. The present study provides an indirect measure of managerial ability of 22 Indian life insurers for the period 2013–14 to 2019–20. For efficiency evaluation of the life insurers both data envelopment analysis and goal programming have been used. In the second stage, the efficiency scores are regressed on several contextual variables (solvency ratio, market share, insurer size and insurer age). The results indicate that the influence of size and insurer age is statistically significant. We have derived estimates of managerial ability for the two methods of evaluation and the results are quite similar. Finally, we have validated the estimates of managerial ability by regressing return on total asset (an important measure of insurer performance) on managerial ability measurers and the regression coefficients are statistically significant.
Ram Pratap Sinha, Bahareh Vaisi
Efficiency Appraisal and Classification of Flexible Random Factors
Abstract
Traditional data envelopment analysis (DEA) models assist in comprehending the performance of decision-making units (DMUs) in situations where a definitive set of inputs and outputs is available. However, in certain real-world scenarios, it becomes necessary to assess the performance of DMUs when dealing with flexible and random measures. Hence, in this chapter, we propose the utilization of two distinct methodologies, namely oriented and non-oriented chance-constrained DEA-based approaches, to evaluate the efficiency of diverse organizations incorporating both stochastic and flexible elements. The chance-constrained DEA-based patterns are transformed into deterministic mixed integer programming problems. A case study is used to estimate the relative efficiency and categorize flexible random variables by utilizing the proposed approach. According to the empirical evidence, efficiency levels exhibit no increment as risk levels increase. The utilization of stochastic observations is more informative when compared to the attainment of efficiency scores through precise measures.
Monireh Jahani Sayyad Noveiri, Sohrab Kordrostami
Performance Evaluation of Indian Banking Financial Sector by Using DEA Approach
Abstract
New approaches that give data beyond what can be obtained from a formal analysis of financial statements are needed to measure and evaluate the effectiveness of the bank’s operations. A mathematical programming approach called Data Envelopment Analysis (DEA) identifies inefficient banks while considering the breadth of the services offered and the resources required to supply these banking services. Based on cross-sectional data from the top 10 nationalized banks for the period of April 2018–April 2022, this study aims to evaluate intra-sector performance in the Indian banking system. This study uses a DEA method to assess the relative effectiveness of a group of Indian banks in the real world. Finally, the bank is ranked using CRS and VRS models based on input and output financial ratio values. The investigation will highlight the banks that function effectively, areas where bank efficiency is not up to par, and probable causes of inefficiency.
R. Venugopal, C. Veeramani, V. T. Dhanaraj, E. Kungumaraj
Application of Data Envelopment Analysis in Decision Making of Civil Engineering Problems
Abstract
Data envelopment analysis (DEA) can convert fractional linear efficiency into a linear programming model for decision analysis. DEA evaluates decision-making units (DMUs) using a nonparametric method. It is a mathematical programming method for assessing DMU performance. This method does not identify general relationships between all units. DEA is a frontier-based statistical regression method. In this chapter, two examples of DEA applications in civil engineering are provided. Two of the applications relate to structural design engineering and earthquake engineering, respectively.
Sanaz Razmyan, Azad Yazdani
A Robust Optimization Approach for Estimating the Most Productive Scale Size in Uncertain Data Envelopment Analysis
Abstract
Productivity measurement, ranking, performance assessment, and benchmarking of homogeneous decision-making units (DMUs), as well as estimating the most productive scale size (MPSS) and returns to scale (RTS), are among the most significant challenges faced by managers and decision-makers (DMs) in various real-world situations and problems. Data envelopment analysis (DEA), a widely used and powerful approach, can be applied to address all of these challenges. It is important to note that, in the presence of uncertain and ambiguous data, traditional DEA approaches cannot be utilized effectively. Therefore, this research presents a novel method for estimating the most productive scale size in DEA within a deep uncertainty environment. Notably, the robust optimization (RO) method, recognized for its effectiveness and practicality in uncertain programming, is employed to propose the robust most productive scale size (RMPSS) model. The implementation of the developed robust MPSS method is demonstrated through a numerical example and a real-life case study from the Iranian healthcare system. Furthermore, the results indicate the effectiveness and efficacy of the proposed approach in estimating MPSS within uncertain DEA.
Pejman Peykani, Farhad Hamidzadeh, Mir Saman Pishvaee, Elaheh Memar-Masjed, Armin Jabbarzadeh
Goal Programming Method for Solving Tri-Level Data Envelopment Analysis
Abstract
Data Envelopment Analysis (DEA) is a methodical approach utilizing linear programming to assess the effectiveness and efficiency of Decision-Making Units (DMUs). It particularly focuses on scenarios where these DMUs possess multiple inputs and outputs. However, when dealing with practical situations, we often come across intricate and interconnected circumstances that operate within a hierarchical framework, with decisions being made in a decentralized manner. The traditional DEA models are not suitable for these hierarchical issues because they assume a single decision-making center with all decision-making power and do not consider the decentralized nature of the decision-making process. Recently, a group of scholars proposed an innovative bi-level programming data envelopment analysis approach to assess the entities within a hierarchical framework, wherein decentralized decision-making involves two tiers of decision-making centers. The bi-level programming DEA takes into account the two levels of decision-making centers and their interactions. It considers the decision-making process as a bi-level optimization problem, where the upper-level decision-making center aims to maximize the overall efficiency of the system, while the lower-level decision-making centers aim to optimize their own efficiencies. In this chapter, the DEA model is extended for units with tri-level structures and then to solve the model, a goal programming approach is applied. Finally, an illustrative example is presented to demonstrate the efficacy of this method.
Morteza Shafiee, Hilda Saleh
Machine Learning Techniques and Efficiency Evaluation: A Survey of Methodological Contributions
Abstract
This chapter surveys the literature on two types of related contributions. The first group is made up of models devoted to adapting well-known machine learning techniques for estimating production frontiers, satisfying shape constraints (free disposability, convexity, …). The second group consists in approaches that apply frontier estimators to classify observations under the classical framework of supervised machine learning with two or more classes. Thus, this survey represents a round-trip between two relatively unconnected fields to date: machine learning and frontier analysis. In particular, we conduct a review of the existing literature devoted to methodological aims.
Juan Aparicio, Miriam Esteve, Qianying Jin
A Literature Review for Nonparametric Frontier Methods Applied to Portfolio Analysis
Abstract
This chapter systematically summarizes previous research on nonparametric frontier methods employed for portfolio performance evaluation and benchmarking, including the diversified and nondiversified frontier models. The former is based on the diversified portfolio frontier, which explicitly considers the diversification effect when combining portfolios, while the latter directly derived from the production involves the envelope of the convex or nonconvex combinations of observed portfolios. Both mainstream methods are reviewed and discussed in single-horizon and multi-horizon portfolio analysis frameworks, respectively, covering the latest developments in this field, existing opportunities, and potential directions for future research. Additionally, we provide a generic production possibility set (PPS) formula for each class of methods to analytically clarify their modeling ideas. This literature study, along with a comprehensive summary of conceptual, methodological, and empirical developments, serves as a reference and guideline for future investigative work on the applications of nonparametric frontier methods in portfolio performance assessment.
Tiantian Ren, Helu Xiao, Zhongbao Zhou
An Optimization Algorithm to Solve Imprecisely Defined Unconstrainted Optimization Problem
Abstract
Many optimal solution driven engineering problems are operated through various parameters which are uncertain in nature. Therefore, for better understanding of the system and estimation of the field variables, the problem can be considered with epistemic type of uncertainties. Here, Trapezoidal Fuzzy Number is considered to visualize the epistemic constants and coefficients. A parametric concept is adopted here to model the governing imprecise algebraic equations to a fuzzy unconstrainted optimization problem. Then a search algorithm is implemented to handle the fuzzy unconstrained minimization problem. Moreover, to quantify the uncertainty, the algorithm with fuzzy theory is discussed in various cases. The different cases are discussed in details through couple of test electrical network problems. Finally, a comparative study of the present approach with others is presented.
Paresh Kumar Panigrahi, Sukanta Nayak
Examining Dimensions and Critical Success Factors of Supply Chains Based on the Blockchain and Internet of Things (B-IoT)
Abstract
The technological revolution in the field of logistics and supply chain brings a large volume of innovations and new challenges. Using today’s powerful and high-speed digital technologies, customers expect faster and on-time ordering and delivery of goods. As a result, organizations are always looking to implement today’s new and transformative technologies to provide services faster and more efficiently. “Internet of Things” and “Blockchain” are expected innovations that will transform the complex supply chain into an integrated and empowered process. IoT innovations such as sensor data and RFID (radio identification technology) provide information to equip us with features such as real-time tracking and alerts to improve decision-making. Blockchain can also increase data security and transparency through decentralized data management. Such data can be turned into critical information to help improve business operations and processes. Therefore, the successful implementation of a supply chain based on these two valuable technologies can be an effective way to improve supply chain processes and help business growth in all supply chain processes. In this research, it has tried to identify and evaluate the most important critical success factors for the implementation of this smart supply chain. Correct understanding and prioritization of these factors can be an effective guide for the powerful implementation of a smart supply chain based on the Internet of Things and Blockchain.
Esmaeil Najafi, Hamed Nozari
Leveraging Data Mining Techniques to Render Unprecedented Opportunities for Business Organizations to Survive and Thrive in the Course of COVID-19
Abstract
Big data in the face of saddling to anticipate, comprehend and react to future occasions becomes an interesting and well-designated platform for the research community associated with some administrative decisions. During a critical situation like the COVID-19 pandemic, to drive the researchers to make different administrative decisions following the different protocols, the existence of a comprehensive review of different data mining techniques to tackle the situation is very lean. In this paper or the methodology corner, we present an audit of methodological advancements of a variety of techniques utilized for mining data which can further render help in analyzing the contemporary issue of industries. We provide insight into how various descriptive and predictive methods of data mining can be leveraged to help different industries to increase their revenue even during these black swan events such as COVID-19.
Pooja Bhakuni, Amrit Das
Optimizing Ecological Development Zone Selection: A Comparative Analysis of AHP and DEA-Modified VAHP Approaches in Geography
Abstract
This chapter delves into the practical application of decision science in geography, specifically focusing on the utilization of the Analytical Hierarchy Process (AHP) methodology. Acknowledging the limitations inherent in the conventional AHP method, a novel approach called Voting Analytical Hierarchy Process (VAHP) is proposed, which seamlessly integrates the principles of Data Envelopment Analysis (DEA). Emphasizing the significance of selecting appropriate ecological development zones, the chapter navigates through the challenges that arise in decision-making within this field. Additionally, the advantages and disadvantages of AHP and DEA are discussed, underscoring their respective roles in multi-criteria decision-making and performance evaluation. The VAHP approach emerges as a potent decision support tool, as it actively engages experts in the prioritization process, mitigating judgmental inconsistencies. Furthermore, the chapter presents a comprehensive comparative analysis of AHP and VAHP, executed through a case study within the realm of geography, thus providing valuable insights into their efficacy and reliability. The outcomes of this research contribute significantly to the advancement of decision science in geography and yield practical implications for promoting sustainable ecological development. As the chapter concludes, it offers insightful recommendations for researchers intending to employ AHP, DEA, and VAHP in analogous decision-making contexts. Overall, this chapter serves as a valuable resource, equipping both academics and practitioners in the field of geography with a comprehensive understanding of the application of AHP, DEA, and VAHP.
Mehdi Soltanifar, Saeid Kamyabi
A Multi-Objective Investment Selection Problem Using Fuzzy and Intuitionistic Fuzzy Approach
Abstract
Investment refers to allocating funds in the expectation of some benefit in the future which in financial terms is the return. In the presence of various investment alternatives, the decision maker stands confused as to which option would offer him the highest return. Thus, the problem of appropriate investment selection can be treated as a multi-objective decision-making (MODM) problem that involves optimization of return on investment, time, risk, etc. However, the time, return, and risk involved in such problems cannot be defined in precise units and are fuzzy in nature. Amidst such a scenario in addition to fuzzy set (FS) theory, intuitionistic fuzzy sets (IFSs) which provide a mathematical framework to deal with imprecise information of the real world can be of much help. It can be seen as an alternative to describe a FS-in situation when the existing data is not enough to define a usual FS. Against this backdrop, this paper is an attempt to develop an IF multi-objective linear model (MOLM) for the investment selection problem where the coefficients of the objective functions (OFs) are represented by IF linear membership (MM) and non-membership (N-MM) functions. The constraints in the problem are treated as crisp. The application of the methodology is explained with the help of a numerical example. Comparing the fuzzy and intuitionistic models it is seen that the IFO model gives optimal results in selection and order allocations to the investors. Also, a Pareto optimality test is performed to test the strength of the solution.
Prabjot Kaur, Nasreen Kausar, Salma Khan, Dragan Pamucar
Revolutionizing Education: An Optimal MAGDM-Based e-Learning Approach for Curriculum Beyond the Classroom
Abstract
The rise of web applications in the form of e-learning websites has presented both opportunities and challenges for academic organizations and individuals involved in education. Many organizations have developed websites to provide education and enhance skills. However, the rapid growth of e-learning brings the challenge of evaluating and selecting the most suitable e-learning websites. One approach to address this challenge is through Multi-Attribute Group Decision-Making (MAGDM) problems. To select the best e-learning website, this study proposes an integrated model in a 2-tuple linguistic q-rung orthopair fuzzy (2TLq-R) set. The proposed model utilizes the 2TLq-R set, a robust mathematical framework capable of handling uncertainties and linguistic assessments in decision-making problems. To improve the model’s effectiveness, new operators for 2TLq-R numbers are introduced, utilizing the Hamacher t-norm (TN) and t-conorm (TCN) operations. These operators enable efficient aggregation and comparison of linguistic evaluations, streamlining the selection process for the e-learning website. With the proposed approach, decision-makers can assess e-learning websites based on multiple attributes, including content quality, interactivity, user experience, and cost-effectiveness. This comprehensive evaluation allows decision-makers to make informed choices when selecting the most suitable e-learning platform. This paper investigates the properties and analyzes specific cases of 2TLq-R operators to understand their characteristics. Moreover, the study combines the Multi-Attributive Border Approximation Area Comparison (MABAC) method with the 2TLq-R operators to incorporate the decision-makers’ psychological behavior. This integration leads to the development of a novel approach called 2TLq-R-MABAC, which effectively addresses real-world MAGDM problems. Additionally, the paper includes an illustrative example and a comparative evaluation to demonstrate the practicality and applicability of the proposed method for selecting the optimal e-learning website.
Sumera Naz, Areej Fatima, Shariq Aziz But, Arooj Ashiq
Balanced Neutrosophic Fermatean Graphs with Applications
Abstract
In the context of ensuring timely and equitable healthcare access in a nation like India, the chapter investigates the concept of balanced neutrosophic fermatean graphs, a specialized form of graph theory that offers a means to establish evenly distributed network connectivity and balanced topologies. The exploration extends to self-complementary and strictly balanced neutrosophic fermatean graphs, showcasing their potential applications in addressing the critical challenge of providing timely medical care to underserved populations within Indian cities. This research not only contributes to the field of graph theory but also presents a structural framework for optimizing medical treatment delivery in government hospitals, with far-reaching implications for improving healthcare outcomes and reducing casualties.
Said Broumi, S. Sivasankar, Assia Bakali, Mohamed Talea
TOPSIS-Based Entropy Measure for N-Valued Neutrosophic Trapezoidal Numbers and Their Application to Multi-Criteria Decision-Making Problems
Abstract
As an extension of the neutrosophic trapezoidal numbers, the N-valued neutrosophic trapezoidal numbers, which are special neutrosophic multisets on real numbers, are used to effectively solve the repetitive uncertainty of decision-makers in multi-criteria decision-making problems. The aim of the chapter is to develop a multi-criteria decision-making method under N-valued neutrosophic trapezoidal numbers. Therefore, we first propose some new generalized distance measures for N-valued neutrosophic trapezoidal numbers. Also, some desirable properties of these measures are shown, and some special cases of the measure are given. We second propose an entropy measure to construct a decision-making method, which is to find the weight of criterias. Then, we develop a multi-criteria decision-making method under N-valued neutrosophic trapezoidal numbers called TOPSIS-based entropy method. Furthermore, we apply it to a decision-making problem to illustrate the effectiveness of the developed method. Finally, a comparative analysis is initiated with the decision-making problem to show the advantages of our method.
Vakkas Uluçay, İrfan Deli
A New Decision-Making Analysis Model Based on the Transformation of Picture Fuzzy Sets into Fuzzy Sets
Abstract
In today's era of data analysis and decision-making, the concept of picture fuzzy sets has emerged as a valuable tool for addressing uncertain real-life problems. Picture fuzzy sets uniquely consider neutral answers and accommodate choices beyond the typical binary yes or no, including the neutral option. However, the application of artificial intelligence, machines, and software is predominantly optimized for traditional fuzzy sets. In this chapter, we propose an innovative approach that transforms picture fuzzy sets into conventional fuzzy sets. This transformation bridges the gap between these two mathematical frameworks, empowering industries to harness the precision of fuzzy logic and control by providing exact fuzzy values for more accurate results. This work contributes to the integration of picture fuzzy sets into the broader landscape of fuzzy logic, enhancing their applicability in various industrial contexts.
Taiwo O. Sangodapo, Nasreen Kausar, Mohammad Y. Chreif
An Enhanced Score Function for Quadripartitioned Single-Valued Neutrosophic Sets
Abstract
The concept of the fuzzy neutrosophic soft matrix (NSM) has advanced by integrating the neutrosophic set (NS) with a soft set to manage uncertain data. The recently introduced theory of fuzzy quadripartitioned NSM generalizes the NSM. Within this quadripartitioned single-valued (QSV) NS, the indeterminacy function’s partition is categorized into contradiction and ignorance values. This study introduces an enhanced score function (SF) and accuracy function (AF) for the QSVNS, addressing existing limitations. Their properties are explored and verified through specific cases. Additionally, an example demonstrates the proposed multi-criteria decision-making method, emphasizing the importance of ranking orders in determining optimal outcomes. A comparative analysis showcases the efficacy of the new SF and AF against traditional methods.
Akanksha Singh, Said Broumi, S. Krishna Prabha, Assia Bakali, Mohamed Talea
Metadata
Title
Analytical Decision Making and Data Envelopment Analysis
Editors
S.A Edalatpanah
Farhad Hosseinzadeh Lotfi
Kristiaan Kerstens
Peter Wanke
Copyright Year
2024
Publisher
Springer Nature Singapore
Electronic ISBN
978-981-9769-72-8
Print ISBN
978-981-9769-71-1
DOI
https://doi.org/10.1007/978-981-97-6972-8

Premium Partner