Genetic Algorithm-Based Movement Patterns for Scrolling-Shooter Games

Chang-Hoon Park, Jinseok Seo

One of the exciting elements of scrolling-shooter games is finding the movement pattern of enemies, and predicting the direction in which they will move to eliminate them.  However, if players identify all the movement patterns of the enemies, they will easily lose interest in game play.  In the end, in today’s game ecosystem, which has a very short life-cycle, players who feel weary will look for new games.  To address this problem, this paper presents a genetic algorithm based technique for dynamically generating movement patterns of enemy aircrafts.  The proposed method evolves the movement patterns designed by a game designer in the early stage to the more complex form of movement according to the player 's ability and vice versa.