The seminar will introduce the topic of autonomous agents and multi-agent systems, which are becoming more and more important in our increasingly interconnected world. Multi-agent systems are generally accepted as valuable tools for designing and building distributed dynamical systems, by using several and often many interacting agents, possibly including humans. In the morning part of the seminar will cover theoretical and technological aspects of agent-based systems with the emphasis on languages, platforms and methods for effectively implementing such systems. We will also discuss the agent-based approach to building model of complex real-world systems. In the second part of the seminar, the general techniques will be illustrated on several concrete application domains where the agent-based approach has already been successfully employed. Applications in air traffic control, network security monitoring and production planning will be presented.

1. Introduction to Agent-based computing (Michal Pěchouček)
   1. Agent-based computing is a subfield of computer science and artificial intelligence, that studies the concepts of autonomy of individual computational processes (running either software applications or hardware robots) and interaction between such heterogeneous autonomous processes.
   2. Agent-based computing leverages results from mathematical logics, game theory, mechanism design, machine learning, automated planning and others.
   3. Multi-agent systems are collections of autonomous agents, which either simulates or control distributed cooperative/competitive systems. Agent technology is a collection of methods, algorithms and software tools that support development of multi-agent systems.
2. Programming Autonomous Agents and Mulit-agent Systems (Peter Novák)
   The session will be dedicated to
   1. An introduction to theoretical foundations of agent-oriented programming and
   2. A brief overview of state-of-the-art programming languages for cognitive agents. We will briefly introduce two main theoretical perspectives on formalization of _intentions_ in the context of the Belief-Desire-Architecture, the currently most followed architectural abstraction for agent oriented programming, and subsequently dedicate attention to selected programming languages for agent programming inspired by the architecture. We will conclude the session with a brief sketch of their extensions towards implementation of agents' social abilities, i.e., multi-agent systems.
3. Agent Platforms (David Šišlák)
   In this session we will give a general description of agent platforms, functions and services platforms provide to agents residing on them and to system developers to speed up the implementation process. In the second part of the session we will present brief overview of existing agent platforms, comparison of their features and performance.
   
4. Agent-based Modelling and Simulation (Michal Jakob)
   The session will introduce the field of (multi-agent)-based modelling and simulation, which is increasingly used for studying complex systems in many areas of human life. In addition to defining the fundamental concepts and methodologies, the session will discuss technical and technological challenges faced when designing and building agent-based simulations.
   
5. Agent Applications in Traffic and Transportation (David Šišlák)
   In this session we will introduce the problem of air traffic control in general, current status in the field and the Free-flight concept. In the second part, we will use the simulation framework to present the solutions of particular technological challenges like path planning and collision avoidance. In the end the large-scale simulations of real-world scenarios will be presented.
   
6. Agent Applications in Network Security Monitoring (Martin Rehák)
   The session will introduce the concept of security monitoring in general and define the problems of system monitoring and autonomous analysis in an uncertain, adversarial environment. Then, using this general theoretical framework, we will use the example of network monitoring and intrusion detection to illustrate the theoretical concepts on a real-world example.
   
7. Agent Applications in Production Planning (Jiří Vokřínek) 
   This session will discuss the applicability of agent based systems in the real industry problems and domains. First the features and benefits of the agent-based planning systems will be presented. Then, we will introduce the agent-based applications for intra-enterprise and extra-enterprise production planning, simulation, and supply chain management developed in cooperation with industry.
    

Registrace na tuto akci již byla ukončena.