Elsevier

Ad Hoc Networks

Volume 107, 1 October 2020, 102222
Ad Hoc Networks

A game theoretic approach for conflict resolution in argumentation enabled social IoT networks

https://doi.org/10.1016/j.adhoc.2020.102222Get rights and content

Abstract

The notion of socially enabled Internet of Things (IoT) termed as Social IoT, emphasized on social interactions among smart objects in an IoT network. Moreover, the concept of speaking and hearing objects further advanced the level of interactions occurring among the IoT devices by empowering them with conversational abilities. Researchers are constantly working to practically realize this vision, keeping in mind the potential benefits of the integration of technologies like argumentation, logical reasoning and artificial intelligence into IoT. Moving along the same lines, this paper investigates the emerging concept of speaking objects and proposes novel reference architecture for argumentation enabled Social IoT systems. The proposed framework lays out a theoretical foundation for development of future applications in the said domain. Further, to overcome the challenges of conflict resolution in argumentation-based decision making, a game theoretic weighted voting scheme is deployed. To ascertain the validity the proposed model, a smart home assisted living use-case scenario is simulated in MATLAB. The simulations are conducted for both weighted and non-weighted voting schemes in a 4-player cooperative game and their performance is evaluated. Experimental results revealed that the proposed scheme outperforms the existing non-weighted voting schemes in terms of throughput which is improved by 35% for quota, q = 0.5. For higher quota values, a minimum improvement of at least 15% is observed. Therefore, the proposed approach confirms to be an effective scheme for decision making in argumentation enabled Social IoT systems.

Introduction

The current trend in IoT is gradually drifting towards social objects from smart objects. Social objects are capable of autonomously establishing social relationships with each other based on some defined principles. The idea of socially enabled objects designated as Social IoT was first build up by Atzori et al. [1]. The ultimate aim of Social IoT is to make the machines (IoT objects) communicate and collaborate with each other in a way that closely resemble with humans. Therefore, the prototypes built to analyze human social behavior and social networks can be utilized to characterize the behavior pattern of social objects as well. Social IoT can be utilized in any sort of IoT-enabled environments such as smart cities, smart homes, smart buildings, assisted living, telemedicine etc. Topical studies in IoT have revealed some additional advantages of Social IoT such as network navigability, object and service discovery, flexibility and trustworthiness [2].

Smart objects are getting smarter with every passing day in light of the escalating computing power embedded to the sensing devices. With the advent of these mounting computing capabilities and advancements in enabling technologies, the sensing devices are now able to process the data at a conceptual level rather than merely sensing and communicating it [3]. These smarter than ever IoT objects are capable of interacting with other objects on their own and undertake best suited action in order to achieve desired goal. Therefore, in the coming future these objects will be able to mimic normal human conversations, dialogues and arguments with the help of advanced enabling technologies such as Artificial Intelligence [4], Logical Reasoning [5], Argumentation [6] etc. Realizing the vision of articulating objects, the authors in [7] introduced the idea of “speaking and hearing objects”. Here, “speaking objects” refer to objects capable of presenting their opinion or arguments as a dialogue in a conversation in order to support a conclusion, communicate their plans, argue or negotiate etc. while “hearing objects” refer to objects capable of taking note of arguments and ideas generated by other speaking objects. Hearing objects tend to be at the receiving end of the conversation. They generally are actuators that work in coalition to achieve a desired goal and plan a course of action to achieve that goal. The combined presence of these speaking and hearing enabled objects in a Social IoT network imparts the network with argumentation and human like conversation abilities.

However, the extensive use of Artificial Intelligence in day-to-day lives might lead to some non-trivial societal and legal implications. The authors in [8], reviewed the social and juristic challenges posed by the replacement of humans with artificially intelligent objects in the coming future. The authors discussed the social dilemma situations in which the machine has to make some ethically questionable decisions. For instance, the dilemma of autonomous vehicles presented in [9], where a driverless vehicle has to choose between saving the vehicle owner by running over some pedestrians or sacrificing the owner’s life but saving the pedestrians. There is no right answer to such dilemmas. In order to protect the society from such inevitable hazards, the authors emphasized on the building of an appropriate regulatory system and a prepared juristic system, which is not only strong enough to handle such challenges but also flexible enough to not restrict technological development. Therefore, it is important to maintain a fine balance between innovation on one hand and legal restrictions on the other.

Out of the several manners in which humans interact with each other, argumentation is undeniably one of the most effective techniques. Arguments generated by smart devices can be utilized to derive appropriate conclusion from the set of possible claims corresponding to a given problem. Argumentation techniques are needed when there is a conflicting situation. To resolve the conflict, no. of arguments in favor of or against the claim can be considered to prove its validity. In the Social IoT domain, argumentation appears to be a useful strategy for decision making in the context of speaking and hearing objects. Sometimes, the decision is to determine which action is needed to be performed in a given situation. Exploring argumentation in context of Social IoT possess notable advantages over conventional interaction methods as presented by authors in [7]. Some of those are discussed further:

  • i.

    Traditional rule based systems need prior determination of the coordination rules. Thus, it is quite hard for these systems to work in unanticipated circumstances. While, argumentation based systems can be surfaced to learn and adapt to work in unforeseen situations [10].

  • ii.

    Decision making in conversation enabled systems, involves a series of arguments or assertions to reach to a conclusion thereby enabling the system to provide explanations to the users [11].

  • iii.

    It enables the human users involved in the system to interact with the system in a more natural way rather than the conventional command based system [12].

Having discussed the effectiveness of argumentation techniques, it will be prudent to integrate them with today’s trending technologies such as IoT and Social IoT. However, for this amalgamation to go smoothly there are still many challenges needed to be overcome. One of the major issues addressed in this paper is reaching the consensus in case of a dispute or disagreement situation occurring in an argumentation based scenario. To resolve this issue, we propose a game theoretic approach for decision making in an argumentation enabled Social IoT network.

Game theory is basically an applied mathematics discipline also known as the rational choice theory. It provides mathematical models for strategic interaction among the parties involved in order to reach a consensus or make some decision. The mathematical models involved in game theory are usually analogous to a game with “win” and “lose” outcomes and the decision makers or participants involved are called the game players [13]. It is typically employed in conflicting situations where no agreement could be reached. Game theory was initially used in the field of economics worldwide. However, with time, its application area is extended to several multi-disciplinary domains such as biology, law, transportation, computer science, psychology, business etc [14]. The business analysts make use of game theoretic principles to model the behaviour patterns of different business organisations [15]. Also, game theory is utilised by the economists to analyse economic competitions, strategic interactions and negotiations etc [16]. Another important application is in the area of politics, where it is widely used in bringing agreement among multiple political parties [17]. Also, game theory based rational decision making approaches are extensively used in legal conflict management [18].

Game theoretic schemes can be chiefly categorized into two basic approaches; cooperative and non-cooperative. In cooperative games, players collectively follow a strategy for the greater good i.e. they cooperate with each other through negotiations and agreements to achieve a common goal. However, in non-cooperative games, the players adopt a strategy for their personal benefit without considering the benefit of others. In computer science, game theory is majorly used to enable automated interaction and coordination among multiple agents (players). In the context of this research work, game theoretic models are deployed for bringing agreement among smart devices possessing different opinions. Also, Weighted Voting Game (WVG) is a renowned cooperative game model extending its applications in computational logic and distributed systems decision making. WVG model provides decision making mechanism in the conflicting situations where multiple parties are involved [19]. It is a simple social choice approach used in many real world problems. The potential benefits of WVG in logical decision making drive the motivation behind this research work. Taking into account all the above concerns, this paper attempts to provide a novel approach for decision making in an argumentation enabled Social IoT networks. Thus, the major contributions of this research article can be characterized into three phases:

  • 1.

    The paper begins with presenting the newly introduced notion of argumentation in context of Social IoT. Further, its characteristics, applications and potential benefits are explored.

  • 2.

    The emerging concept of “speaking” and “hearing” objects in an IoT network is discussed. Also, the state of the art in the said domain is illustrated. Further, a layered architecture for argumentation enabled Social IoT networks is proposed.

  • 3.

    A game theoretic approach for conflict resolution and decision making in proposed system is developed. To establish its validity and effectiveness, the proposed model is simulated in MATLAB and its performance is evaluated.

Rest of the paper is arranged into different sections as follows. Section 2 throws light on the vision of speaking and hearing objects. Section 3 portrays the current state of the art in the domain of Social IoT and argumentation. In Section 4, architecture for argumentation based Social IoT networks is proposed and its various components are discussed. Further, Section 5 highlights the simulation of the proposed model and its performance evaluation. Finally, Section 6 concludes the paper by discussing its major findings and future objectives.

Section snippets

Speaking and hearing objects: the vision

There has been significant development in the field of IoT objects (devices) in the past decade. Earlier IoT objects possessing some sort of sensing, actuating and communicating abilities were regarded as smart objects. Sensor devices had the ability to sense, process and communicate the information while actuator devices could perform some action based on user commands. However, with the growth in the field of Information and Communication Technologies (ICTs), devices were able to hold more

State of the art

The concept of integrating social network and Internet of Things has gained significant attention by researchers and developers over the past decade. This led to the development of Social IoT in its current form. One of the major reasons for its increasing popularity is the potential benefits a Social IoT network offers over an IoT network, such as enhanced network navigability, more flexibility and ease of service discovery [25], [26]. This accounts for the reason for preferring Social IoT

Proposed system model

In this section we describe the proposed model for decision making in argumentation enabled Social IoT networks with the help of a layered architecture (Fig. 2) which is to the best of our knowledge is the first ever proposed system of its kind. The workflow of the system is defined in terms of three major phases it passes through starting with basic data acquisition from sensors to reaching the final conclusion via complex decision making process. The major rationale of the proposed system is

Performance evaluation

In this section, the efficacy of the proposed scheme is confirmed with the help of simulation. We make use of a simulation model, to evaluate the performance of the proposed scheme. Also, to establish the primacy of proposed scheme, it is compared with the existing scheme in terms of throughput. For the performance evaluation, we count on the probability of success of the proposed scheme over existing scheme by simulating it in a use case scenario.

Conclusion and future work

In this paper, the concept of speaking and hearing objects is discussed and architecture for Social IoT systems involving speaking objects is presented. The proposed model addresses two major issues occurring in an argumentation based system; conflict resolution and decision making. In particular, the proposed model had two major objectives; one is building reference architecture for argumentation enabled Social IoT systems and second is enabling the system with automated decision making. It

Conflict of Interest

None.

Ms. Nancy Gulati is working as a Research Scholar in the Department of Computer Science and Engineering, Guru Nanak Dev University, RC, Jalandhar, India. She received her Master’s degree in Computer Science and Engineering from National Institute of Technology, Jalandhar, India working on Resource Constraints in Wireless Sensor Networks. She is currently working as a Senior Research Fellow at Guru Nanak Dev University, Punjab, India. Her research interests covers applications of Internet of

References (55)

  • L. Atzori et al.

    From “smart objects” to “social objects”: the next evolutionary step of the internet of things

    IEEE Commun. Mag.

    (2014)
  • L. Atzori et al.

    SIoT: Giving a social structure to the internet of things

    IEEE Commun. Lett.

    (2011)
  • M. Endler et al.

    An approach for real-time stream reasoning for the internet of things

    2017 IEEE 11th International Conference on Semantic Computing (ICSC)

    (2017)
  • A.I. Maarala et al.

    Semantic data provisioning and reasoning for the internet of things

    2014 International Conference on the Internet of Things (IOT)

    (2014)
  • M. Lippi et al.

    An argumentation-based perspective over the social IoT

    IEEE Internet Things J.

    (2018)
  • M. Perc et al.

    Social and juristic challenges of artificial intelligence

    Palgrave Commun.

    (2019)
  • J.-F. Bonnefon et al.

    The social dilemma of autonomous vehicles

    Science

    (2016)
  • L. Bottou

    From machine learning to machine reasoning

    Mach. Learn.

    (2014)
  • A. Fiedler et al.

    Argumentation in explanations to logical problems

    International Conference on Computational Science

    (2001)
  • E.I. Sklar et al.

    Argumentation-based dialogue games for shared control in human-robot systems

    J. Hum.-Rob. Interact.

    (2015)
  • S. Kim

    Game Theory Applications in Network Design

    (2014)
  • B.A. Bhuiyan

    An overview of game theory and some applications

    Philos. Progress

    (2016)
  • I.K. Geckil et al.

    Applied Game Theory and Strategic Behavior

    (2016)
  • M. Kandori

    Evolutionary game theory in economics

    Adv. Econ. Econom.

    (1997)
  • P.D. Straffin Jr

    Power and stability in politics

    Handb. Game Theory Econ. Appl.

    (1994)
  • D.G. Baird et al.

    Game Theory and the Law

    (1998)
  • Y. Bachrach et al.

    Analyzing power in weighted voting games with super-increasing weights

    International Symposium on Algorithmic Game Theory

    (2016)
  • Cited by (0)

    Ms. Nancy Gulati is working as a Research Scholar in the Department of Computer Science and Engineering, Guru Nanak Dev University, RC, Jalandhar, India. She received her Master’s degree in Computer Science and Engineering from National Institute of Technology, Jalandhar, India working on Resource Constraints in Wireless Sensor Networks. She is currently working as a Senior Research Fellow at Guru Nanak Dev University, Punjab, India. Her research interests covers applications of Internet of Things (IoT) and its enabling technologies. Moreover, she is working on integration of Social IoT and healthcare technologies.

    Dr. Pankaj Deep Kaur is working as an Assistant Professor in the Department of Computer Science and Engineering, Guru Nanak Dev University, RC, Jalandhar, India. She received her Bachelor’s Degree in Computer Applications (2000) and Master’s Degree in Information Technology (2003) from Guru Nanak Dev University, Amritsar, India. She completed her Ph.D. in Resource Scheduling in Cloud Computing from Thapar University, Patiala (2014) and has over ten years of teaching and research experience. Her research interests include Cloud Computing and Big data

    View full text