Advanced Studies in Multi-Criteria Decision Making

Sarah Ben Amor, Adiel Teixeira de Almeida, Joao Luis de Miranda, Emel Aktas, Sarah Ben Amor, Adiel Teixeira de Almeida, Joao Luis de Miranda, Emel Aktas

• Publisher: Chapman and Hall/CRC
• Year: 2019
• ISBN:
  9781351722100
• Edition: 1
• Topic: Commerce
• Subtopic: Prise de décision

With contributions from some of the top academics and scientists in the field, Advanced Studies in Multi-Criteria Decision Making presents an updated view of the landscape of Decision Sciences, current research topics, the interaction with other sciences and fields, as well as the prospects and challenges at an international level.

Given that Decision Sciences are recognized today as indispensable for confronting the major societal challenges in science and technology, this book would be of interest to decision-makers, managers, and researchers from academia, and industrial/services companies that would like a fresh insight into MCDM.

Features

• 
  
  

• Integrates a wide range of scientific fields with a general reader approach, including applied researchers from the social, business, enterprise sciences

• 
  
  

• Suitable for academics and professionals

• 
  
  

• Presents a broad coverage of MCDM tools either in industry or in services companies and systems

• 
  
  

• Provides a fresh overview on MCDM studies promoted by prestigious R&D institutions

CHAPTER 1

Implications of World Mega Trends for MCDM Research

Hannele Wallenius and Jyrki Wallenius

CONTENTS

1.1 Introduction

1.2 Internet Searches

1.3 Big Data (and Artificial Intelligence)

1.4 The Sharing (or Platform) Economy

1.5 Climate Change, Concern for Environment

1.6 How Is MCDM Changing?

References

1.1 INTRODUCTION

Digital technology is making rapid advances. The implications for people, companies, and societies are pervasive. It is difficult to foresee all the changes these developments will cause. Understandably, most individuals, many businesses and government leaders are not aware of, let alone prepared for the future changes. According to Brechbuhl from Dartmouth College, this ignorance was the driver behind the recent report, Deep Shift: Technology Tipping Points and Societal Impact, of the World Economic Forum.

The envisioned changes will bring about (1) digital connectivity, independent of time and place, and (2) tools for quickly analyzing vast amounts of digital data. In the World Economic Forum’s report, the changes are grouped into six “mega-trends.” We borrow freely from the report.

1. The Internet—world’s access to the Internet will continue improving; people’s interaction with it will become more ubiquitous

2. Further enhancements in computing power, communications technologies, and data storage, and the ability to interface with digital technology, anytime using multiple devices

3. The “Internet of Things”

4. Big data and Artificial Intelligence (AI)—the ability to access and analyze huge amounts of data; coupled with the “ability” of computers to make decisions based on this data

5. The sharing (or platform) economy and distributed trust (based on, for example, the block chain technology)

6. 3D-printing

These trends will greatly impact our lives, businesses, and governments—even universities—all around the world. As the World Economic Forum’s Report astutely observes, our lives are increasingly being driven and enabled by software. The envisioned changes will be so profound and rapid that large segments of societies have difficulty in keeping up with the developments as users of technology.

The potential of the digital technology is huge, both in enhancing traditional industrial processes (robotics), and even more importantly in generating novel digital services. Many aspects of health care are also benefiting tremendously from new technologies. The digital revolution has begun, although decades (centuries) are needed for its full potential to be realized. One interesting cause of the Internet and social media (which totalitarian governments try to control) is the increased transparency of societies, which helps to improve democracy.

Besides technology mega-trends, there are other highly important mega-trends. These mega-trends, unlike technology mega-trends, are generally perceived as challenges or threats to humankind. Some of them are discussed in PwCForesight#megatrends and by the World Economic Forum:

1. Demographic and social change taking place in many countries (aging populations, decreasing fertility, urbanization, refugee problem)

2. Increasing world population: growing need for food, clean water, and cheap energy

3. Climate change, concern for environment

The mega-trends, whether technology related or non-technology related, pose real concerns, challenges, or even threats to humankind. Most certainly, all of these mega-trends force governments and businesses to operate more efficiently under resource scarcity. Regarding technology mega-trends, privacy issues and security issues are not easy to solve, and today’s societies are grappling with them. Moreover, with robots/AI “outsmarting” many individuals (with time, perhaps most individuals), what do most people do in year 2118? Brechbuhl asks the good question, “What will happen to the sense of worth, place, and contribution to society that human beings have derived from work throughout much of recorded history?” To make matters worse, who guarantees that the AI-driven robots are (programmed to be) friendly toward humankind?¹

We choose technology mega-trends 1, 4, and 5, and non-technology mega-trend 3 from the World Economic Forum’s list, for a closer look. What role can multi-criteria decision making (MCDM) play in them? How can MCDM help? What MCDM concepts will be useful? Recall that our lives are increasingly being driven and enabled by software. We think that it is a good starting point that many MCDM scholars can write their own software. Hence, we should be able to provide tools, software, and ideas to capitalize on rising opportunities and tackle problems resulting from the world’s mega-trends.

1.2 INTERNET SEARCHES

E-commerce is continuing to transform commerce. To an increasing extent people make purchases online. Surprisingly (to us), besides travel and leisure industries, the clothing or fashion industry is almost driving the change. Typically when people buy online, they use some search engines, such as Google. It is not uncommon that the cheapest products or services emerge on top of the list. A typical example is flight tickets between two cities. Incidentally, this apparently is forcing airlines to adopt the strategy originally followed by low-cost airlines of charging extra for better seats, meals, baggage, etc. One problem is that the search engines are not good enough in differentiating among offers (what they actually contain and how much customers value if a bag or meal is included in the price). MCDM scholars could develop better search engines! Search engines, which would not only be based on price, but other attributes as well. Keyword searches have their limitations.

Because of the abundance of offerings online, whether movies, music, or restaurant ads, many companies (and academics) have found it worthwhile to develop so called recommender systems. A recommender system is a subclass of information filtering systems that seeks to predict the “rating” or “preference” that a user would give to an item (Wikipedia). Recommender systems have become increasingly popular in recent years and are extensively used, for example, in choosing what movies to watch, what music to listen to, what news to watch, which books to read, and which restaurants to visit.

The underlying logic in recommender systems can be categorized into collaborative-filtering approaches and content-based–filtering approaches (Waila et al., 2016). Collaborative-filtering approaches are based on the idea of building a model from a user’s past behavior as well as other users’ behavior (items previously purchased). The logic of incorporating other person’s likes is that if other people found this item (or similar items) popular, so would you! Content-based–filtering approaches develop a set of characteristics that an item possesses (which you liked) to recommend additional items with similar properties.

Consumers generally appreciate recommender systems. However, we hesitate recommending them to filter news items that one sees. If an individual is solely or largely dependent on reading news in social media, as opposed to traditional media, recommended (filtered) by a system, the set of news offered becomes narrow, representing a very narrow worldview. We think that in such cases, the recommender systems should periodically suggest different types of news, to broaden the person’s horizon! (Of course, we are assuming that a broader horizon would be better than a narrower one.) But what such news would be, and how to do it, may not be trivial. It seems that Facebook CEO Mark Zuckerberg’s ideas are different regarding the development of Facebook. In a recent interview by CNBC Business News and Finance, he says that Facebook will change its algorithm so that users will see less public content from businesses or publishers and more posts from their friends.

The logic underlying recommender systems should be understandable to MCDM scholars, although such systems have traditionally been developed and studied by computer scientists and AI scholars. We urge MCDM scholars to develop better recommender systems. Both MCDM and recommender systems are about modeling user’s preferences (Lakiotaki et al., 2011).

Voting advice Applications (VAAs) are online systems to help voters find worthy candidates to vote for in national, presidential, and regional elections. Such VAAs are highly popular in many European countries, where sometimes more than half of the electorate use them. They are based on both the candidates and the voters answering a set of questions concerning political preferences. The system (the algorithm) then finds the candidates and party, which are “closest” to the voter’s political preferences. The development of such VAAs involves solving many MCDM/behavioral decision-making problems. The questions must be discriminating, and there cannot be too many of them. They must have proper Likert-scales to make distance measurement meaningful. What distance measure should one use? Are the questions of equal importance to voters or should importance weights be used? If yes, how are they determined? Are voters interested in voting for candidates who have a higher likelihood of becoming elected?

Jyrki Wallenius (2017) gave a keynote on this topic at the Ottawa MCDM Conference. They also have a paper detailing the development of their VAAs and its implementation in Finland (Pajala et al., 2018). We urge other scholars to further work on their respective country’s popular VAAs. It is an important problem, and in particular, in multi-party, multi-candidate elections, voters benefit from the use of such support provided by VAAs by making them much more aware of what the candidates stand for.

1.3 BIG DATA (AND ARTIFICIAL INTELLIGENCE)

According to a recent issue by The Economist, companies’ most valuable resource is data. Data is being continuously generated from various sources, including cash registers, mobile phones, and Internet sites visited by millions of people daily. There is a realization by the corporate world that they should better use this data to their (strategic) advantage.

Typical advertising and marketing agencies or departments do not know how to analyze big data, even though they realize its importance or potential. The need for people possessing analytics skills is high. What role does big data play in advertising? In a nutshell, big data can be used to help create targeted and personalized campaigns that increase the efficiency of advertising or marketing. How is this done? Simply by gathering information and learning about user behavior. Many reward and loyalty programs are based on the use of consumer data. Recommender systems use past purchases or searches to make new recommendations. An interesting phenomenon is the use of social media by ad agencies. It is easy to document and share experiences as customer or consumer in social media. It is not uncommon that thousands of people read these posted reviews and are influenced by them. The world of social media offers interesting research opportunities to help businesses but also to understand human social behavior (Ghosh et al., 2017).

Another area where big data will find its uses is medicine or health care. Various monitoring instruments continuously generate data, as do human genome studies. They eventually lead to better preventive and actual care and more accurate diagnostics. An interesting problem from the perspective of MCDM is how to better incorporate patients’ views on their own healthcare plans and treatment decisions. A more general level concern in health care is to make the system more efficient and more personalized. Healthcare decisions naturally have to deal with multiple criteria, and complex tradeoffs between cost, the quality of care, and even potential loss of lives. Wojtek Michalowski’s (University of Ottawa) work is a good example of the type of impactful work a person with an Operations Research/MCDM background can do in health care. Jack Kitts (2017), President and CEO of Ottawa Hospital, gave a keynote at the Ottawa MCDM Conference, in part, based on Michalowski’s collaboration with the hospital.

AI is a tremendously important field today. Part of the work uses Kohonen’s neural nets (Kohonen, 1988). The idea is to build learning “robots,” which could eventually make decisions on behalf of humans. An example is self-driving automobiles. Such “robots” need to be programmed to follow certain rules. They must make complex moral choices as well. Work is also currently being conducted to incorporate emotions into “robots.” We ask, whose emotions? Our personal view is that we would hesitate to delegate decision-making powers in important matters to “robots,” no matter how “intelligent” they are. We feel that humans should be in control of their own lives. AI is a good tool, but a dangerous master—something the ancient people said of fire.

1.4 THE SHARING (OR PLATFORM) ECONOMY

According to Wikipedia, sharing economy is an umbrella term with a range of meanings and is often used to describe economic activity involving online transactions. It grew out of the open-source community and referred to peer-to-peer–based sharing of resources and access to goods and services. The term is often used in a broader sense to describe sales transactions conducted via online market places (platforms). Online auctions are an example of such a mar...

Table of contents

• Cover
• Half Title
• Series Page
• Title Page
• Copyright Page
• Table of Contents
• Foreword
• Preface
• Editors
• Contributors
• CHAPTER 1 ■ Implications of World Mega Trends for MCDM Research
• CHAPTER 2 ■ MCDA/M in Telecommunication Networks: Challenges and Trends
• CHAPTER 3 ■ SISTI: A Multicriteria Approach to Structure Complex Decision Problems
• CHAPTER 4 ■ Applying Intangible Criteria in Multiple-Criteria Optimization Problems: Challenges and Solutions
• CHAPTER 5 ■ Some Methods and Algorithms for Constructing Smart-City Rankings
• CHAPTER 6 ■ Agricultural Supply Chains Prioritization for Development of Affected Areas by the Colombian Conflict
• CHAPTER 7 ■ Decision Making and Robust Optimization for Medicines Shortages in Pharmaceutical Supply Chains
• CHAPTER 8 ■ Using Spatial Decision Models for Rank Ordering Chocolates
• CHAPTER 9 ■ Multi-Criteria Decision Planning with Anticipatory Networks to Ensuring the Sustainability of a Digital Knowledge Platform
• CHAPTER 10 ■ A Robust Approach for Course of Action Comparison and Selection in Operation Planning Process
• CHAPTER 11 ■ Analyzing the Relationship between Human Development and Competitiveness Using DEA and Cluster Analysis
• INDEX