Sejarah, Teori Dasar dan Penerapan Reinforcement Learning: Sebuah Tinjauan Pustaka

Jeky Andreanus, Ade Kurniawan

Abstract


Today's research on the topic of Machine learning has increased sharply. Machine learning is the future of the world, in the future it will be a revolution in every computer-bound science. This paper examines the field of Reinforcement learning from a computer science perspective. Reinforcement learning is part of Machine learning. In general machine learning is divided into three categories, namely supervised learning, unsupervised learning, and reinforcement learning. Supervised learning requires labeled data to analyze data, training and make conclusions, which can be used for mapping new values. Otherwise, Unsupervised learning does not use labeled data, which is more suitable for relatively irregular problems. In contrast to Reinforcement learning that is based on trial and error, by experimenting on the environment then get a response that will improve its ability. This work is summarized based on the history of Reinforcement learning and the selection of current research. This paper discusses central issues in reinforcement learning, from history, Reinforcement learning models, multiinform Reinforcement learning, including comparison of exploration and exploitation. Ends with a Reinforcement learning implementation survey of several systems. Reinforcement learning is the most suitable Machine learning in learning new things from scratch without human intervention in learning, most of Reinforcement learning is used for in-game learning. But the learning may  takes a long time and is uncertain.

Dewasa ini penelitian mengenai topik Machine learning telah meningkat tajam.  Machine learning adalah masa depan dunia, kedepannya ini akan menjadi revolusi dalam segalah ilmu yang terikat dengan komputerisasi. Paper ini meneliti bidang Reinforcement learning dari perspektif ilmu komputer. Reinforcement learning merupakan bagian dari Machine learning. Secara umum machine learning dibagi menjadi tiga kategori, yaitu supervised learning, unsupervised learning, dan reinforcement learning. Supervised learning memerlukan data berlabel untuk menganalisis data, pelatihan dan membuat kesimpulan, yang dapat digunakan untuk pemetaan nilai-nilai baru. Sebaliknya, Unsupervised learning tidak memenggunakan data berlabel, yang mana lebih cocok untuk masalah yang relatif tidak beraturan. Berbeda dengan Reinforcement learning yang berbasis trial and error, dengan mencoba-coba pada lingkungannya kemudian mendapatkan respon yang akan meningkatkan kemampuannya. Karya ini dirangkum berdasarkan sejarah bidang Reinforcement learning dan pemilihan riset saat ini. Paper ini membahas isu-isu sentral dalam reinforcement learning, mulai dari sejarah, model Reinforcement learning, multiagent Reinforcement learning termasuk melakukan perbandingan dari eksplorasi dan eksploitasi. Diakhiri dengan survei penerapan Reinforcement Learning terhadap beberapa sistem. Reinforcement learning merupakan Machine learning yang paling cocok dalam mempelajari hal baru dari nol tanpa campur tangan manusia dalam pembelajarannya, kebanyakan dari Reinforcement learning digunakan untuk belajar dalam game. Namun pembelajaran yang dilakukan membutuhkan waktu lama dan tidak pasti.


Keywords


reinforcement learning, sejarah, trial and error, model RL, multiagent RL, eksplorasi dan eksploitasi, tinjauan pustaka

Full Text:

PDF

References


K. Arulkumaran, M. P. Deisenroth, M. Brundage, and A. A. Bharath, “Deep reinforcement learning: A brief survey,” IEEE Signal Process. Mag., vol. 34, no. 5, 2017.

A. Kurniawan, I. Riadi, and A. Luthfi, “Forensic Analysis and Prevent of Cross Site Scripting in Single Victim Attack Using Open Web Application Security Project (Owasp) Framework,” J. Theor. Appl. Inf. Technol., vol. 95, no. 6, pp. 1363–1371, 2017.

Cristina and A. Kurniawan, “Sejarah , Penerapan , dan Analisis Resiko dari Neural Network : Sebuah Tinjauan Pustaka,” vol. 03, no. 02, pp. 259–270, 2018.

Q. He, N. Li, W. J. Luo, and Z. Z. Shi, “A survey of machine learning algorithms for big data,” Moshi Shibie yu Rengong Zhineng/Pattern Recognit. Artif. Intell., vol. 27, no. 4, pp. 327–336, 2014.

J. Qiu, Q. Wu, G. Ding, Y. Xu, and S. Feng, “A survey of machine learning for big data processing,” EURASIP J. Adv. Signal Process., vol. 2016, no. 1, 2016.

E. Alpaydin, Introduction to Machine Learning. London: The MIT Press, 2014.

J. Jesson, L. Matheson, and F. M. Lacey, Doing Your Literature Review :Traditional and Sistematic Techniques. 2011.

J. L. Galvan and M. C. Galvan, Writing literature reviews: A guide for students of social and behavioral sciences. 2017.

R. S. Sutton and a G. Barto, “Reinforcement learning: an introduction.,” IEEE Trans. Neural Netw., vol. 9, no. 5, pp. 127–144, 1998.

N. D. Nguyen, T. Nguyen, and S. Nahavandi, “System Design Perspective for Human-Level Agents Using Deep Reinforcement Learning: A Survey,” IEEE Access, vol. PP, no. 99, p. 1, 2017.

R. Giryes and M. Elad, “Reinforcement Learning: A Survey,” Eur. Signal Process. Conf., pp. 1–2, 2011.

J. Schmidhuber, “A possibility for implementing curiosity and boredom in model-building neural controllers,” Proc. Int. Conf. Simul. Adapt. Behav., pp. 222–227, 1991.

L. Busoniu, R. Babuska, B. De Schutter, and B. De Schutter, “A comprehensive survey of multiagent reinforcement learning,” Syst. Man, Cybern. Part C Appl. Rev., vol. 38, no. 2, pp. 156–172, 2008.

E. Amadou, O. Diallo, A. Sugiyama, and T. Sugawara, “Learning to Coordinate with Deep Reinforcement Learning in Doubles Pong Game,” 2017.

A. Pastore, U. Esposito, and E. Vasilaki, “Modelling Stock-market Investors as Reinforcement Learning Agents,” 2015.

J. Won, “Stock price prediction using reinforcement learning,” pp. 690–695, 2001.

J. Xin, H. Zhao, and D. Liu, “Application of Deep Reinforcement Learning in Mobile Robot Path Planning,” no. 16.

D. Schwung, F. Csaplar, A. Schwung, and S. X. Ding, “An application of reinforcement learning algorithms to industrial multi-robot stations for cooperative handling operation,” Proc. - 2017 IEEE 15th Int. Conf. Ind. Informatics, INDIN 2017, pp. 194–199, 2017.


Refbacks



Copyright (c) 2018 Jurnal Telematika

Creative Commons License
This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License.

____________________________________________

Institut Teknologi Harapan Bangsa

Jl. Dipatiukur no. 80-84 Lt. 2

Bandung 40132


Creative Commons License

Jurnal Telematika is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.