Adaptive Scenario Selection in Serious Games Using Finite State Machines and Agent-Based Models

Abstract

This study presents a hybrid approach to scenario adaptation in serious games, using Finite State Machines (FSMs) and Agent-Based Models (ABMs). Focusing on the educational RPG genre of software development, the proposed model aims to automatically adjust the behavior of non-playable characters (NPCs) and the game's progression based on the player's actions and preferences. The methodology included a literature review, followed by the development of a simulation in the JFLAP software, integrating FSMs to manage states and ABMs to promote dynamic and realistic reactions. The results highlight the viability of this integration, which offers an immersive experience by combining predictability and flexibility in the behavior of NPCs. As a limitation, we identified the need for more advanced tools to integrate the technologies. For future work, it is proposed that frameworks be developed to facilitate this approach's implementation.