Pengujian Parameter Algoritma Genetika dan Feed-Forward Neural Networks pada Permainan Ular Klasik

AHMAD BISRY; CECEP MUHAMAD SIDIK RAMDANI; SITI YULIYANTI

doi:10.26760/mindjournal.v9i2.135-152

Pengujian Parameter Algoritma Genetika dan Feed-Forward Neural Networks pada Permainan Ular Klasik

AHMAD BISRY, CECEP MUHAMAD SIDIK RAMDANI, SITI YULIYANTI

Sari

Abstrak

Konfigurasi parameter yang tepat sangat penting untuk memaksimalkan kinerja dari sebuah algoritma. Algoritma genetika dan neural networks memerlukan pemilihan parameter yang sesuai dalam penggunaannya. Pada permainan ular, performa diukur dari score dan efisiensi runtime. Penelitian ini menguji parameter untuk menemukan konfigurasi optimal bagi kedua algoritma. Permainan ular digunakan sebagai model eksperimen karena metrik kinerja yang jelas, seperti score yang didapat dan beberapa rintangan yang ada. Sebanyak 60 eksperimen dilakukan untuk membandingkan jumlah generasi dan populasi, mutation chance, dan jumlah neuron pada hidden layer. Hasil penelitian menunjukkan konfigurasi dengan generasi lebih besar dari populasi adalah yang paling optimal, menghasilkan score setara dengan generasi dan populasi yang sama besar, namun dengan runtime lebih efisien. Mutation chance 0.1% merupakan yang terbaik dibandingkan dengan 0.2% sampai 0.5%. Selain itu, hidden layer dengan 16 neuron lebih efisien dibandingkan 24 neuron, baik dari segi score maupun runtime.

Kata kunci: Algoritma genetika, neural networks, Permainan ular klasik

Abstract

Appropriate parameter configuration is crucial to maximizing algorithm performance. Genetic algorithms and neural networks require careful parameter selection. In the game of Snake, performance is measured by score and runtime efficiency. This research tests parameters to find optimal configurations for both algorithms. Snake serves as an experimental model due to clear performance metrics such as score and various obstacles. Sixty experiments compare generation and population sizes, mutation chances, and neuron counts in hidden layers. Findings indicate that configurations with larger generations than populations are optimal, yielding scores similar to equal-sized generations and populations but with more efficient runtime. A 0.1% mutation chance outperforms rates of 0.2% to 0.5%. A hidden layer with 16 neurons proves more efficient than 24 neurons in both score and runtime aspects.

Keywords: Genetic algorithm, neural networks, classic snake game

Teks Lengkap:

PDF

Referensi

Boris, T., & Goran, S. (2017). Evolving neural network to play game 2048. 24th Telecommunications Forum, TELFOR 2016. https://doi.org/10.1109/TELFOR.2016.7818911

Brown, J. A., de Araujo, L. J. P., & Grichshenko, A. (2021). Snakes AI Competition 2020 and 2021 Report. http://arxiv.org/abs/2108.05136

Carr, J. (2014). An Introduction to Genetic Algorithms. Senior Project (Vol. 1, no. 40, pp. 7).

Chi Yuen, M., Wuan Yeong, L., Chen Yi Kang, E., Qaisar Syed, S., &

Arabee Abdul Salam, Z. (2021). Investigating parameters of genetic algorithm and neural network on classic snake game. Journal of Applied Technology and Innovation, 5(2).

Hamdia, K. M., Zhuang, X., & Rabczuk, T. (2021). An efficient optimization approach for designing machine learning models based on genetic algorithm. Neural Computing and Applications, 33(6), 1923–1933. https://doi.org/10.1007/s00521-020-05035-x

Hau Hor, S., Jeh Tan, S., Kye Yan, M., Arabee bin Abdul Salam, Z., & Shin Sim, Y. (2022). Snake Game: A genetic neural network approach. In Journal of Applied Technology and Innovation, 6(1).

Kong, S., & Mayans, J. A. (2021). Automated Snake Game Solvers via AI Search Algorithms.

Kumar, R., Memoria, M., Gupta, A., & Awasthi, M. (2021). Critical Analysis of Genetic Algorithm under Crossover and Mutation Rate. Proceedings - 2021 3rd International Conference on Advances in Computing, Communication Control and Networking, ICAC3N 2021, (pp. 976–980). https://doi.org/10.1109/ICAC3N53548.2021.9725640

Ma, Y. (2024). Optimization of basic PID control algorithm based on genetic algorithm and Matlab. Theoretical and Natural Science, 30(1), 178–186. https://doi.org/10.54254/2753-8818/30/20241103

Miller, M., Washburn, M., & Khosmood, F. (2019). Evolving unsupervised neural networks for Slither.io. ACM International Conference Proceeding Series. https://doi.org/10.1145/3337722.3341837

Mishra, Y., Kumawat, V., & Selvakumar, K. (2019). Performance analysis of flappy bird playing agent using neural network and genetic algorithm. Communications in Computer and Information Science, 1025 CCIS, (pp. 253–265). https://doi.org/10.1007/978-981-15-1384-8_21

Rahul Ramesh Patil. (2023). AI-Infused Algorithmic Trading: Genetic Algorithms and Machine Learning in High-Frequency Trading. International Journal For Multidisciplinary Research, 5(5). https://doi.org/10.36948/IJFMR.2023.V05I05.5752

Shiruru, K. (2016). An Introduction To Artificial Neural Network. https://www.researchgate.net/publication/319903816

Uthansakul, P., Anchuen, P., Uthansakul, M., & Khan, A. A. (2020). QoE-Aware Self-Tuning of Service Priority Factor for Resource Allocation Optimization in LTE Networks. IEEE Transactions on Vehicular Technology, 69(1), 887–900. https://doi.org/10.1109/TVT.2019.2952568

Wei, Z., Wang, D., Zhang, M., Tan, A. H., Miao, C., & Zhou, Y. (2018). Autonomous agents in snake game via deep reinforcement learning. Proceedings - 2018 IEEE International Conference on Agents, ICA 2018, (pp. 20–25). https://doi.org/10.1109/AGENTS.2018.8460004

Zhang, R., & Cai, R. (2020). Train a snake with reinforcement learning algorithms. https://openreview.net/forum?id=iu2XOJ45cxo

DOI: https://doi.org/10.26760/mindjournal.v9i2.135-152