ABSTRACT
The development of IoT technologies raises the question of their self-optimization. For this purpose we propose an approach based on a multi-agent interpretation, modeling discrete space and time by Cellular Automata (CA), and applying a game-theoretical model known as spatial Prisoner's Dilemma (SPD). A 2D space is occupied by agents which belong to species with certain strategies. Each agent participates in games with neighbors and its goal is to maximize its cumulated payoff. Species competing for space apply locally a mechanism of evolutionary selection, where the cumulated payoff is considered as fitness. As a result of the competition, more profitable species replace less profitable ones. While agents act locally to maximize their incomes we study conditions of emerging collective behavior measured by the global average total payoff of which the players are not aware. We show that collective behavior based on achieving in a fully distributed way a Nash equilibrium (NE) can emerge if some conditions of the game are fulfilled, in particular when an "income sharing mechanism" is introduced.
- C. Cox. 2012. An Introduction to LTE: LTE, LTE-Advanced, SAE and 4G Mobile Communications. John Wiley & Sons Ltd, Chapter 19, 291--299. Google ScholarCross Ref
- Y. Ishida and Y. Katsumata. 2008. A Note on Space-Time Interplay through Generosity in a Membrane Formation with Spatial Prisoner's Dilemma. In Knowledge-Based Intelligent Information and Engineering Systems, KES 2008 (LNCS), I. Lovrek, R. J. Howlett, and L. C. Jain (Eds.), Vol. 5179. Springer, 448--455. Google ScholarDigital Library
- Y. Khaluf, E. Ferrante, P. Simoens, and C. Huepe. 2017. Scale invariance in natural and artificial collective systems: a review. Journal of The Royal Society Interface 14, 136 (2017), 20170662. Google ScholarCross Ref
- N. Msadek and T. Ungerer. 2017. Trustworthy self-optimization for organic computing environments using multiple simultaneous requests. J. Syst. Archit. 75 (2017), 26--34. Google ScholarDigital Library
- A. Nedić and R. M. May. 2018. Distributed Optimization for Control. Annual Review of Control, Robotics, and Autonomous Systems 1, 1 (2018), 77--103. Google ScholarCross Ref
- M. A. Nowak and J. Liu. 1992. Evolutionary games and spatial chaos. Nature 359 (1992), 826--829.Google ScholarCross Ref
- M. Osborne. 2009. An Introduction to Game Theory. Oxford University Press.Google Scholar
- P. Östberg, J. Byrne, and et al. 2017. Reliable capacity provisioning for distributed cloud/edge/fog computing applications. In 2017 European Conference on Networks and Communications, EuCNC 2017,. IEEE, 1--6. Google ScholarCross Ref
- F. Seredyński and J. Gąsior. 2020. Behavior Optimization in Large Distributed Systems Modeled by Cellular Automata. In Parallel Problem Solving from Nature, PPSN 2020, (LNCS), T. Bäck, M. Preuss, and et al (Eds.), Vol. 12269. Springer, 678--690. Google ScholarDigital Library
- F. Seredyński, J. Gąsior, and R. Hoffmann. 2020. The Second Order CA-Based Multi-agent Systems with Income Sharing. In Cellular Automata, ACRI 2020, (LNCS), T. M. Gwizdalla, L. Manzoni, and et al (Eds.), Vol. 12599. Springer, 134--145. Google ScholarDigital Library
- W. Trumler, A. Pietzowski, B. Satzger, and T. Ungerer. 2007. Adaptive Self-optimization in Distributed Dynamic Environments. In Proceedings of the First International Conference on Self-Adaptive and Self-Organizing Systems, SASO 2007. IEEE Computer Society, 320--323. Google ScholarDigital Library
Index Terms
- Towards evolutionary self-optimization of large multi-agent systems
Recommendations
Competition and Cooperation Mechanisms for Collective Behavior in Large Multi-agent Systems
Computational Science – ICCS 2022AbstractWe consider a 2-dimensional discrete space modeled by Cellular Automata consisting of cells that can be occupied by agents. There exist several types of agents which differ in their way of behavior related to their own strategy when they ...
Behavior Optimization in Large Distributed Systems Modeled by Cellular Automata
Parallel Problem Solving from Nature – PPSN XVIAbstractWe consider a distributed system modeled by the second-order Cellular Automata (CA) and interpreted as a multi-agent system, where interactions between agents are defined by a spatial Prisoner’s Dilemma game. The idea of the second-order CA is ...
The Second Order CA-Based Multi-agent Systems with Income Sharing
Cellular AutomataAbstractWe consider a multi-agent system composed of the second-order nonuniform Cellular Automata (CA)–based agents, where a spatial Prisoner’s Dilemma (PD) game describes the interaction between agents. Each agent has some strategy that can change in ...
Comments