Efficiency bounds in computations on open multi-agent systems
Files
Monnoyer_16211300_2018.pdf
Open access - Adobe PDF
- 1.37 MB
Details
- Supervisors
- Faculty
- Degree label
- Abstract
- Multi-agent systems are systems constituted of multiple independent and interactive entities acting towards some objective. Really widespread for their advantages in solving big problems in a decentralized way, it is getting needed to extend them to the case where arrivals and departures of agents happen. This case arises as the size of the systems grows, and through the study of some problems that naturally present such a property (such as computer failures in a network). Such systems are called Open Multi-Agent Systems, and represent a really young theory in need for development. A step in the understanding of such a theory consists in establishing algorithms and analyses for basic problems, that can be the base of more complex tasks. In this Master thesis, we intend to determine fundamental performance limitations on basic consensus problems in the framework of Open Multi-Agent Systems, under the form of bounds that can be used as quality criteria for algorithms solving these problems.