Modern and Interdisciplinary Problems in Network Science

A Translational Research Perspective

Zengqiang Chen, Matthias Dehmer, Frank Emmert-Streib, Yongtang Shi

• Publisher: CRC Press
• Year: 2018
• Print ISBN: 9780815376583
• eBook ISBN: 9781351237284
• Topic: Mathematics
• Subtopic: Mathematics General
• Index: Mathematics

Modern and Interdisciplinary Problems in Network Science: A Translational Research Perspective covers a broad range of concepts and methods, with a strong emphasis on interdisciplinarity. The topics range from analyzing mathematical properties of network-based methods to applying them to application areas. By covering this broad range of topics, the book aims to fill a gap in the contemporary literature in disciplines such as physics, applied mathematics and information sciences.

Chapter 1

The Spread of Strategies Predicted by Artificial Intelligence in Networks

Chunyan Zhang

Contents

1.1Individual Intelligence in Collective Behaviors

1.2Strategy Competition Driven by Particle Swarm Optimization

1.2.1Model settings

1.2.2Simulation study

1.3Strategy Spreading Controlled by Fuzzy Neural Network

1.3.1Problem statement

1.3.2Framework of model

1.3.3Dynamics of strategy competition

1.3.3.1Effect of evaluation level and verification

1.3.3.2Microscopic behaviors of individuals

1.3.3.3Formation of C-clusters

1.4Conclusion

1.5Glossary

References

1.1Individual Intelligence in Collective Behaviors

Recently, the individual and collective behaviors in multi-agent systems have received much attention of researchers from many areas, including engineering, biology, sociology, and so on. Among the related topics, the competition among strategies which indicate different benefits for the players has become the focus.

In fact, cooperation among uncorrelated individuals is necessary in order to allow the common group to offer significant advantages for them. The basic conflicts lie on the fact that, in general, the involved individuals are self-centered according to the definition. However, many examples can verify the existence of cooperative behavior in nature. For example, animals collaborate in families to raise their offspring, or in foraging groups to prey or to defend against predators. In social society, cooperation among unrelated agents will be beneficial for raising a more advanced society. For example, the altruistic cooperation among members in society can help to achieve shared goals and making efficient use of the common resource [1].

Nowadays, a growing number of researchers take an interest in evolutionary game theory, as it provides an effective framework for describing the strategy competition in several scenarios. Different with the traditional model with the assumption of full rational players and complete knowledge about the game, evolutionary models assume that agents can choose their strategies by a trial-and-error learning process. In this case, they can gradually find the strategies with better performance than others. Through this type of strategy updating in repeated games, strategies with worse performance tend to be weeded out in the system.

In this sense, as the objects of the study, the characteristics of the involved agents deserves significant attention, as the reasonable modeling about the individuals will help us to approximate to the real systems. Only in this way, the collective behaviors based on them can get better understanding. It is understandable that agents involved in the strategy competition and collective actions have (simple or complicated, the extent may be heterogeneous) intelligence. According to the introduction in artificial intelligence, an intelligent agent can be seen as an autonomous entity. She observes the variation in her surroundings and makes her action toward achieving goals. To achieve their goals, agents with intelligence may also learn or use knowledge about others or surroundings. The described intelligent agents here are closely related to agents in economics, engineering, sociology, as well as in many interdisciplinary modeling and computer simulations.

Thus based on our understanding, the hypothesis of simple agents should be relaxed, as the studied are the real agents involved in strategy competition in real social systems. Individuals here may present their intelligence in the following...

Table of contents

• Cover
• Half Title
• Title Page
• Copyright Page
• Contents
• About the Editors
• List of Editors
• Contributors
• 1. The Spread of Strategies Predicted by Artificial Intelligence in Networks
• 2. The Spread of Multiple Strategies in the Complex Networks
• 3. The Epidemic Spreading Processes in Complex Networks
• 4. Measurements for Investigating Complex Networks
• 5. Overview of Social Media Content and Network Analysis
• 6. Analysis of Critical Infrastructure Network
• 7. Evolving Networks and Their Vulnerabilities
• 8. Review of Structures and Dynamics of Economic Complex Networks: Large Scale Payment Network of Estonia
• 9. Predicting Macroeconomic Variables Using Financial Networks Properties
• 10. Anomaly Detection in Dynamic Complex Networks
• 11. Finding Justice through Network Analysis
• Index