|
|
Simulation of Entities Collective Behavior in Virtual World
Vymazal, Tomáš ; Žák, Pavel (referee) ; Láník, Aleš (advisor)
Theme of this work is to evaluate and compare aviable paradigms for entity control in virtual worlds, and to implement one of these paradigms in application. Dynamic finite state machine upgraded using genetic algorithms has been chosen. This paradigm should make agent's behavior better and adapt agent to required state: i.e. make agent harvest resources in virtual world. Output of this work is application for running evolution and application for 3D view of agent's behavior.
|
|
|
Training Intelligent Agents in Unity Game Engine
Vaculík, Jan ; Chlubna, Tomáš (referee) ; Matýšek, Michal (advisor)
The goal of this work is to design applications, which demonstrate the power of machine learning in video games. To achieve this goal, this work uses the ML-Agents toolkit, which allows the creation of intelligent agents in the Unity Game Engine. Furthermore, a series of experiments showing the properties and flexibility of intelligent agents in several real-time scenarios is presented. To train the agents, the toolkit uses reinforcement learning and imitation learning algorithms.
|
|
|
Simulation of Negotiation and Argumentation Protocols
Říha, Michal ; Samek, Jan (referee) ; Zbořil, František (advisor)
This work deals with communication in multiagent systems. The protocols for negotiation and argumentation are shown, and model example of their usage is described. We describe hierarchical model of trust in contexts, that is used for representation of agent's believes. The argumentation protocol for those agents is designed, and is used for solving conflicts.
|
|
|
Multirobot Path Planning in a Dynamic System
Dokoupil, Ladislav ; Veigend, Petr (referee) ; Zbořil, František (advisor)
This thesis deals with the problem of dynamic environment search using multi-agent systems. The primary result of this work is participation in the MAPC2022 contest, but the application can be found in the exploration of unknown space, assuming finite visibility and unlimited distance of communication of the agents. After describing the current methods for solving the given problem and their limitations, an algorithm based on ant colony optimization is proposed. Graphs were then created with data from running program with various parameters. The result of work is agents synchronization improvements and overall optimization of the platform involved in the mentioned contest from previous year. As a result half more of explored space was measured compared to previous solution.
|
|
|
Multirobot Path Planning in a Dynamic System
Dokoupil, Ladislav ; Veigend, Petr (referee) ; Zbořil, František (advisor)
This thesis deals with the problem of dynamic environment search using multi-agent systems. The primary result of this work is participation in the MAPC2022 contest, but the application can be found in the exploration of unknown space, assuming finite visibility and unlimited distance of communication of the agents. After describing the current methods for solving the given problem and their limitations, an algorithm based on ant colony optimization is proposed. Graphs were then created with data from running program with various parameters. The result of work is agents synchronization improvements and overall optimization of the platform involved in the mentioned contest from previous year. As a result half more of explored space was measured compared to previous solution.
|
|
|
Training Intelligent Agents in Unity Game Engine
Vaculík, Jan ; Chlubna, Tomáš (referee) ; Matýšek, Michal (advisor)
The goal of this work is to design applications, which demonstrate the power of machine learning in video games. To achieve this goal, this work uses the ML-Agents toolkit, which allows the creation of intelligent agents in the Unity Game Engine. Furthermore, a series of experiments showing the properties and flexibility of intelligent agents in several real-time scenarios is presented. To train the agents, the toolkit uses reinforcement learning and imitation learning algorithms.
|
|
|
Simulation of Entities Collective Behavior in Virtual World
Vymazal, Tomáš ; Žák, Pavel (referee) ; Láník, Aleš (advisor)
Theme of this work is to evaluate and compare aviable paradigms for entity control in virtual worlds, and to implement one of these paradigms in application. Dynamic finite state machine upgraded using genetic algorithms has been chosen. This paradigm should make agent's behavior better and adapt agent to required state: i.e. make agent harvest resources in virtual world. Output of this work is application for running evolution and application for 3D view of agent's behavior.
|
|
|
Simulation of Negotiation and Argumentation Protocols
Říha, Michal ; Samek, Jan (referee) ; Zbořil, František (advisor)
This work deals with communication in multiagent systems. The protocols for negotiation and argumentation are shown, and model example of their usage is described. We describe hierarchical model of trust in contexts, that is used for representation of agent's believes. The argumentation protocol for those agents is designed, and is used for solving conflicts.
|
|
|
Legal Status of Police Agents: Study 5.293
Veselá, Hana
Práce se zabývá použitím institutu agenta ve vybraných zemích a jeho zakotvením v právním řádu. Srovnávány jsou evropské úpravy (agent kontrolor) s odlišnou úpravou ve Spojených státech (agent provokatér). Zvláštní část práce se zabývá relevantními rozsudky Evropského soudu pro lidská práva. Použití skrytého (tajného) vyšetřovatele - agenta se změněnou identitou patří mezi zvláštní vyšetřovací postupy, které podléhají zpřísněné kontrole. Jejich použití totiž zasahuje do soukromí osob, které jsou pomocí této metody vyšetřovány. Mezi zvláštní vyšetřovací prostředky patří nejenom použití agenta se změněnou identitou, ale také odposlechy, odposlechy telefonních hovorů, sledování dat v elektronickém styku (včetně jejich rozšifrování), a pod.
Fulltext:  PDF
|
|
|
Incorporation of Agent Oriented Approaches to the Complex Process Description Methodology
Smolík, Jan ; Řepa, Václav (advisor) ; Vojtáš, Peter (referee) ; Bukovský, Ivo (referee)
The main objective of this thesis is to integrate agent oriented concepts into a complex methodology for description of business processes MMABP that is being developed on University of Economics, Prague. The first part describes and explains agent oriented approaches to business process modeling (i*/TROPOS, AOR, OOEM and UFO ontology) and demonstrates them on a case study. MMABP methodology is then compared with these approaches and evaluated. The evaluation concludes that MMABP is incomplete in relation to agent oriented concepts of plan, plan execution, desire, intention, commitment and claim and also to concepts of stable and unstable state. Concepts of goal, agent and business process are evaluated as inaccurate. Second part of the theses arguments and defines MMABP extension that eliminates identified inadequacies. Metamodel is amended with the concept of plan execution which is intentionally executed action that leads to a goal fulfillment. Plan is a description of this special kind of action. Agent is an entity that has goals and is capable of executing plans. Business process received a new definition and is defined as something that defines plans (as a prescription of actions), has goals and uses agents to execute actions. Business process is not defined as a specialization of a complex event anymore. The thesis also specifies procedure where concepts are step-by-step transformed from one diagram to another.
|
guest ::
