National Repository of Grey Literature 2 records found  Search took 0.01 seconds. 
Robust multi-agent path finding
Nekvinda, Michal ; Barták, Roman (advisor) ; Pilát, Martin (referee)
The thesis is devoted to finding robust non-conflict paths in multi-agent path finding (MAPF). We propose several new techniques for the construction of these types of paths and describe their properties. We deal with the use of contingency planning and we create a tree plan for the agents where the specific path is chosen by the agents during the execution based on the current delay. Next we present an algorithm that increases robustness while maintaining the original length of the solution and we combine it with the previous approach. Then we will focus on the method of increasing robustness by changing the speed of agents. Finally we experimentally verify the applicability of these techniques on different types of graphs. We will show that all the proposed methods are significantly more robust than the classic solution and they also have certain advantages over previously known constructions of robust plans.
Artificial intelligence and game strategy in Carcassonne board game
Nekvinda, Michal ; Holan, Tomáš (advisor) ; Kuboň, David (referee)
The bachelor paper deals with the development of artificial intelligence for the Carcassonne board game for two to five players. Different approaches to creating a suitable game strategy are presented. Then, selected strategies are tested. Most of the paper is dedicated to the development by means of genetic algorithms. The paper presents detailed results of the comparison of all implemented artificial intelligences. In addition, it explains how the most successful decision-making method works. The most advanced artificial intelligences are capable of defeating human players. Additionally, the game environment for this game was implemented. It provides a graphical user interface application that offers a human-to-computer gameplay and a simple console application to perform the artificial intelligence performance tests.

See also: similar author names
2 Nekvinda, Matěj
Interested in being notified about new results for this query?
Subscribe to the RSS feed.