Realistické chování postav a zvířat ve hrách

Abstract

This bachelor’s thesis focuses on the design and implementation of realistic behavior for non-player characters (NPCs), with an emphasis on animal entities in the Unreal Engine 5 environment. The output of the thesis includes an analysis of game artificial intelligence principles, particularly decision-making models such as behavior trees and utility AI, navigation systems, and advanced animation techniques. Furthermore, the thesis presents an application demonstrating the control of NPC characters, where entities respond to internal states (health, hunger, stamina) and external stimuli through a hybrid decision-making system. The work also includes the implementation of multiplayer replication and a save system, along with an evaluation of the chosen approaches in terms of performance and realism.