Open AccessDynamic game difficulty balancing in real time using Evolutionary Fuzzy Cognitive Maps with automatic calibration
Author(s) -
Lizeth Joseline Fuentes Perez,
Luciano Arnaldo Romero Calla,
Anselmo Montenegro,
Luis Valente,
Esteban Clua
Publication year - 2016
Publication title -
journal on interactive systems
Language(s) - English
Resource type - Journals
ISSN - 2763-7719
DOI - 10.5753/jis.2016.668
Subject(s) - fuzzy cognitive map , computer science , fuzzy logic , representation (politics) , artificial intelligence , flexibility (engineering) , robotics , graph , calibration , machine learning , process (computing) , knowledge representation and reasoning , fuzzy classification , fuzzy control system , robot , theoretical computer science , mathematics , statistics , politics , political science , law , operating system
Fuzzy Cognitive Maps (FCM) is a paradigm used to represent knowledge in a simple and concise way, expressing the grade of relation that exists between concepts and causal relationships. Due to its flexibility, FCM has been successfully applied in numerous applications in diverse research fields, such as, robotics, medical diagnosis, decision problems in information technology, games, and so forth. However, one critical drawback is the determination of the weights in the representation graph, which is generally done by an expert. The present paper proposes a semi-automated method for calibrating the weights in a solution for the problem of dynamic game difficulty balancing (DGB) using Evolutionary Fuzzy Cognitive Maps (E-FCM). The proposed algorithm adjusts the weights in real time, ensuring an equilibrium between the values generated according to the expert’s contribution (based on a static analysis) and the changes produced in the values of the concepts by the calibration process during the simulation (a dynamic analysis).