Utilization of the Particle Swam Optimization Algorithm in Game Dota 2

Hendrawan Armanto (Universitas Negeri Malang Institut Sains dan Teknologi Terpadu Surabaya)
Harits Ar Rosyid (Universitas Negeri Malang)
Muladi Muladi (Universitas Negeri Malang)
Gunawan Gunawan (Institut Sains dan Teknologi Terpadu Surabaya)

Register: Jurnal Ilmiah Teknologi Sistem Informasi
Vol 10 No 2 (2024): July

Dota 2, a Multiplayer Online Battle Arena game, is widely popular among gamers, with many attempting to create efficient artificial intelligence that can play like a human. However, current AI technology still falls short in some areas, despite some AI models being able to play decently. To address this issue, researchers continue to explore ways to enhance AI performance in Dota 2. This study focuses on the process of developing artificial intelligence code in Dota 2 and integrating the particle swarm optimization algorithm into Dota 2 Team's Desire. Although particle swarm optimization is an old evolutionary algorithm, it is still considered effective in achieving optimal solutions. The study found that PSO significantly improved the AI Team's Desire and enabled it to win against Default AI of similar levels or players with low MMR. However, it was still unable to defeat opponents with higher AI levels. Furthermore, this study is expected to assist other researchers in developing artificial intelligence in Dota 2, as the complexity of the development process lies not only in AI but also in language, structure, and communication between files.