The Modified Flocking Algorithm for Optimizing Non Player Character Movements

Abstract

Non player character is an important part of a game that can make gameplay more interesting and challenging. That reason also makes it important to study, especially for non player characters that are clustered in large numbers where programming of such non player characters is relatively difficult to do. Quite a lot of games with non player characters in the form of birds that are clustered but on average the game does not apply good movement so that the motion of birds is relatively slow, monotonous and unrealistic. In this research, we will try to modify the Flocking Algorithm, as one of the popular algorithms for non player character groups, so that birds can fly scattered in terrain while maintaining flocking. In its implementation, the new algorithm was successfully designed and inserted in flocking, which was then applied to a 3D game. The test results are quite satisfactory with only 6.6% collisions on 100 agents with a maximum frame rate of 10 fps. Non player characters succeeded well toward 3 targets in less than 2 minutes and the results of beta testing were an average of 81% of respondents "VERY AGREE" The implementation of the Flocking Algorithm and its modification worked well and resulted in a realistic movement of non player characters.