Adversarial Reinforcement Learning for Procedural Content Generation

Abstract This work presents a Procedural Content Generation (PCG) method based on a Neural Network Reinforcement Learning (RL) approach that generates new environments for Virtual Reality (VR) learning applications. The primary objective of PCG methods is to algorithmically generate new content (e.g., environments, levels) in order to improve user experience. Researchers have started exploring the integration of Machine Learning (ML) algorithms into their PCG methods. These ML approaches help explore the design space and generate new content more efficiently. The capability to provide users with new content has great potential for learning applications. However, these ML algorithms require large datasets to train their generative models. In contrast, RL based methods take advantage of simulation to train their models. Moreover, even though VR has become an emerging technology to engage users, there have been few studies that explore PCG for learning purposes and fewer in the context of VR. Considering these limitations, this work presents a method that generates new VR environments by training an RL agent using a simulation platform. This PCG method has the potential to maintain users’ engagement over time by presenting them with new environments in VR learning applications.

Download Full-text

Procedural Content Generation: Better Benchmarks for Transfer Reinforcement Learning

10.1109/cog52621.2021.9619000 ◽

2021 ◽

Author(s):

Matthias Muller-Brockhausen ◽

Mike Preuss ◽

Aske Plaat

Keyword(s):

Reinforcement Learning ◽

Procedural Content Generation ◽

Content Generation

Download Full-text

Deep Reinforcement Learning for Procedural Content Generation of 3D Virtual Environments

Journal of Computing and Information Science in Engineering ◽

10.1115/1.4046293 ◽

2020 ◽

Vol 20 (5) ◽

Author(s):

Christian E. López ◽

James Cunningham ◽

Omar Ashour ◽

Conrad S. Tucker

Keyword(s):

Reinforcement Learning ◽

Virtual Environments ◽

A Priori ◽

Three Dimensional ◽

Generative Models ◽

Primary Objective ◽

Training Data ◽

Procedural Content Generation ◽

3D Virtual Environments ◽

Content Generation

Abstract This work presents a deep reinforcement learning (DRL) approach for procedural content generation (PCG) to automatically generate three-dimensional (3D) virtual environments that users can interact with. The primary objective of PCG methods is to algorithmically generate new content in order to improve user experience. Researchers have started exploring the use of machine learning (ML) methods to generate content. However, these approaches frequently implement supervised ML algorithms that require initial datasets to train their generative models. In contrast, RL algorithms do not require training data to be collected a priori since they take advantage of simulation to train their models. Considering the advantages of RL algorithms, this work presents a method that generates new 3D virtual environments by training an RL agent using a 3D simulation platform. This work extends the authors’ previous work and presents the results of a case study that supports the capability of the proposed method to generate new 3D virtual environments. The ability to automatically generate new content has the potential to maintain users’ engagement in a wide variety of applications such as virtual reality applications for education and training, and engineering conceptual design.

Download Full-text

Island Generator pada Game Open world Menggunakan Algoritma Perlin noise

Repositor ◽

10.22219/repositor.v2i7.601 ◽

2020 ◽

Vol 2 (7) ◽

pp. 965

Author(s):

Naufal Azzmi ◽

Lailatul Husniah ◽

Ali Sofyan Kholimi

Keyword(s):

Performance Testing ◽

Game Development ◽

Procedural Content Generation ◽

Open World ◽

Game World ◽

Perlin Noise ◽

And Performance ◽

Game Developers ◽

Pcg Method ◽

Content Generation

AbstrakPerkembangan game pada saat ini berkembang dengan sangat cepat, dalam perkermbangan game topik AI adalah topik yang paling banyak diteliti oleh beberapa peneliti khususnya pada pembuatan suatu konten game menggunakan metode PCG (procedural content generation). Pada pembuatan sebuah game world menggunakan metode PCG sudah banyak developer game yang sukses dengan mengimplementasikan metode ini, metode ini banyak digunkan pada geme dengan genre RPG, Rouglikes, Platformer, SandBox, Simulation dan lain sebagainya, Pada penelitian ini berfokus pada pengembangan sebuah game world generator untuk game berjenis open world yang berupa sebuah kepulauan dengan metode PCG dengan menggunakan algoritma perlin noise sebagai algoritma pembentuk textur utama pulau yang dimana pada penelitian ini memanfaatkan beberapa variable noise seperti octave, presistance dan lacunarity guna untuk menambah kontrol dari hasil textur yang dihasilkan serta algoritma penempatan pulau untuk membuat sebuah game world yang menyerupai sebuah kepulauan. Dari hasil uji generator terkait degan pengujian playability dan performa dapat disimpulkan bahwa generator yang dikembangkan playable serta performa yang dianaliasa menggunakan notasi Big O menunjukkan (linear). Abstract Game development is currently growing very fast, game development AI is the most discussed topic by most researchers especially in the developing of game content using the PCG (procedural content generation) method. In making a game world using the PCG method, many game developers have succeeded by implementing this method, this method is widely used on RPGs, Rouglikes, Platformers, SandBox, Simulations and ect,. This study focuses on developing a game world generator game for open world type games in the form of an archipelago using the PCG method using the noise perlin algorithm as the island's main texturizing algorithm which in this study utilizes several noise variables such as octave, presistance and use for add control of the texture results as well as the island placement algorithm’s to create a game world that resembles an archipelago form. From the generator test results related to the playability and performance testing, it shows that map are being generated by the generators are playable and performance that are analyzed using Big O notation show O (n) (linear).

Download Full-text