Procedural Content Generation Research Articles

Procedural Content Generation in Games: A Survey with Insights on Emerging LLM Integration

Procedural Content Generation (PCG) is defined as the automatic creation of game content using algorithms. PCG has a long history in both the game industry and the academic world. It can increase player engagement and ease the work of game designers. While recent advances in deep learning approaches in PCG have enabled researchers and practitioners to create more sophisticated content, it is the arrival of Large Language Models (LLMs) that truly disrupted the trajectory of PCG advancement. This survey explores the differences between various algorithms used for PCG, including search-based methods, machine learning-based methods, other frequently used methods (e.g., noise functions), and the newcomer, LLMs. We also provide a detailed discussion on combined methods. Furthermore, we compare these methods based on the type of content they generate and the publication dates of their respective papers. Finally, we identify gaps in the existing academic work and suggest possible directions for future research.

Making New Connections: LLMs as Puzzle Generators for the New York Times' Connections Word Game

The Connections puzzle is a word association game published daily by The New York Times (NYT). In this game, players are asked to find groups of four words that are connected by a common theme. While solving a given Connections puzzle requires both semantic knowledge and abstract reasoning, generating novel puzzles additionally requires a form of metacognition: generators must be able to accurately model the downstream reasoning of potential solvers. In this paper, we investigate the ability of the GPT family of Large Language Models (LLMs) to generate challenging and creative word games for human players. We start with an analysis of the word game Connections and the unique challenges it poses as a Procedural Content Generation (PCG) domain. We then propose a method for generating Connections puzzles using LLMs by adapting a Tree of Thoughts (ToT) prompting approach. We evaluate this method by conducting a user study, asking human players to compare AI- against human-generated Connections puzzles. Our findings show that LLMs are capable puzzle creators, and can generate diverse sets of enjoyable, challenging, and creative Connections puzzles as judged by human users.

Level2image: A Utility for Making 2D Tile-Based Level Images with Overlays

2D tile-based levels are a common format for video game research, particularly in procedural content generation. Often, tiles are represented as text characters. Here, we describe level2image, a utility that provides a flexible means for converting such levels into a variety of image formats. It can use text tiles or substitute image tiles or backgrounds. There is support for various geometry overlays, such as paths through the level, areas of interest, or boundaries. The utility is a Python script, intended to reduce duplicated work creating such converters within the game research community.

Using EPCG for Designing a Hexagon Tangram Puzzle

Designing and tuning a game is a complex creative process, with a variety of tools have been designed for assisting human designers in this process. But, many existing tools either entirely take the design away from humans, or they only assist in generating content - they cannot be used to create an entire game from scratch. This paper explores how Exhaustive Procedural Content Generation (EPCG) can be integrated as a co-creative partner in the design process. We describe how EPCG was used alongside human designers to create a novel Hexagon Tangram puzzle, from initial conception to puzzle curriculum. EPCG is used to answer design questions at each step of the design, which enable human designers to make more informed design choices.

Toward Space-Time WaveFunctionCollapse for Level and Solution Generation

WaveFunctionCollapse (WFC) is a constraint-satisfaction-based approach to procedural content generation via machine learning (PCGML). It is relatively easy to implement and requires very little training data, making it a popular approach. Generated game levels are guaranteed to look locally similar totheir example tilemaps; however, local adjacency rules often fail to capture global solvability rules, potentially making many such levels unplayable. Existing approaches to improving the solvability of WFC-generated levels typically require adding additional game-specific information in the form of global constraints, substantially increasing the complexity and time required for setup. The purpose of this work is to explore whether using level solutions as training data can allow WFC to learn solvability constraints and game mechanics. We have implemented a novel space-time approach that uses three-dimensional space-time blocks representing solutions to 2D levels as both input and output. Experiments using this method show that space-time WFC is capable of demonstrating localized game mechanics and creating small playable levels with given solutions. However, levels are slow to generate, and some high-level constraints are still not captured.

Design and implementation of a survival game towards achieving user’s satisfaction by providing balanced difficulty

Recently due to high popularity of video games, Procedural Content Generation (PCG) becomes an approach for automatically creating levels in game with balanced difficulty. The PCG is a solution towards automatic creation of levels with balanced difficulty. Among the existing works, not only very few recent PCG-based difficulty balancing solutions is found, but algorithms are also computationally intensive due to the presence of complex techniques like evolutionary computing, graph algorithms etc. In this paper, we develop a low overhead mathematical model based PCG scheme with the rule of variation and curve fitting. Firstly, relationship between difficulty and performance parameters of the game is setup using this model. Then, to make the model more reliable, difficulty parameters are adjusted to cope with the real performance of a player. Finally, the proposed scheme is tested on a survival game ‘Helip’ developed by us. The game shows satisfactory performance in response time, user friendliness, visual aesthetics and innovation based on opinion of the players. Moreover, the proposed difficulty balancing scheme is evaluated in terms of opinion score and player’s performance to be more balanced compared to state-of-the-art techniques. Our balancing scheme also has the lowest overhead in terms of running time.

Enhancing 3D Printing with Procedural Generation and STL Formatting Using Python

Three-dimensional printing has become a fast-growing industry. The first phase of this technology is the design of a 3D object to personalize it and optimize its production. This paper explores the procedural generation of the 3D model. The article aims to present the method of procedurally generating 3D objects in Python. Procedural content generation is the automated creation of content using algorithms. Most often, as part of procedural generation, a small number of input parameters and pseudo-random processes are used to generate content that will meet the requirements. The programming techniques for object customization in Python optimize the manufacturing process. Moreover, procedural generation speeds up the model design, and if developers use 3D scanning methods and artificial intelligence, production can be personalized, which is in line with the concept of Industry 4.0.

Comparative Analysis of Metaheuristic Algorithms for Procedural Race Track Generation in Games

Procedural Content Generation (PCG) aims to automatically generate the content of games using algorithmic approaches, as this can reduce the cost of game design and development. PCG algorithms can be applied to all elements of a game, including terrain, maps, stories, dialogues, quests, and characters. A wide variety of search algorithms can be applied to PCG problems; however, those most often used are variations of evolutionary algorithms. This study focuses on comparing three metaheuristic approaches applied to racetrack games, with the specific goal of evaluating the effectiveness of different algorithms in producing game content. To that end, a Genetic Algorithm (GA), Artificial Bee Colony (ABC), and Particle Swarm Optimization (PSO) are applied to a game-level design task to attempt to identify any discernible differences in their performance and identify whether alternative algorithms offer desirable performance characteristics. The results of the study indicate that both the ABC and PSO approaches offer potential advantages to Genetic Algorithm implementation.

Procedural content generation in pervasive games: state of affairs, mistakes, and successes

PurposeIn ubiquitous games, sensor data and the player’s dynamic profile personalize the experience, adapting scenarios and difficulty to the player’s context. However, manually creating custom virtual environments becomes impractical due to the scalability required in the real world, where gameplay depends on the player’s environment and various points of interest. Procedural content generation (PCG) naturally emerges as an automated solution. This study aims to review the state of the art among pervasive games that use procedural generation techniques.Design/methodology/approachThe authors performed a systematic review of the literature on PCG techniques in ubiquitous games. The research methodology included initial database searches, forward snowballing and backward snowballing.FindingsThe authors selected 40 articles published from 2008 to 2022 after analyzing 1,017 PCG articles. Many of these paper use dynamic difficulty generation methods that affect game customization, balance and playability. The majority of titles are in the area of education and require rigorous evaluation. Some articles mention the use of PCG but provide limited details of algorithms or processes.Research limitations/implicationsThe low presence of recent articles stands out, which can lead to two conclusions: poor use of indexing of articles in the area or low amount of research over these years. One of the search bases did not enter due to the number of keywords used.Practical implicationsThe paper is a guide for researchers in the area of pervasive games who are interested in using PCG techniques in their games.Originality/valueThis approach revealed articles combining PCG and ubiquitous games, with no previous systematic reviews at this intersection.

A Review on DISDAIN: An Auto Content Generation VR Game

This document discusses the manner in which DISDAIN has been conceptualized, developed, implemented, and evaluated, looking into its major mechanics, algorithm- based content creation as well as user interaction and all impact it has made on virtual reality gaming. It explores DISDAIN’s design and implementation with much detail to highlight how procedural content generation can improve replayability and immersion in virtual reality games. The results from user testing and gameplay analysis indicate that the game has been successful in providing various experiences that are interesting while at the same time ensuring high levels of player contentment. Besides, this study analyzes technical performance and adaptive AI of DISDAIN so as to project what may be expected from dynamic content generation in VR gaming in future. In general, these results indicate that DISDAIN is a starting point for new developments in VR gaming henceforth while at the same time setting new standards for this kind of game category

A mixed-initiative design framework for procedural content generation using reinforcement learning

Currently, there are a significant and growing number of games and players. Creating digital games becomes a challenging task, as manual game development is costly and time-consuming. A technique known as procedural content generation (PCG) can potentially reduce both the time and production costs of games. It is feasible to automate the creation process by utilizing artificial intelligence techniques and PCG, assisting game designers in their tasks. PCG is not a novel concept, and there is a diverse range of algorithms aimed at automatically generating content in games. However, a significant number of these techniques do not incorporate artificial intelligence. This paper introduces the PCGRLPuzzle framework used to generate procedural scenarios through reinforcement learning agents trained with the policy proximal optimization algorithm. The process of building scenarios poses a challenging problem due to the existence of an exponential number of possibilities. The framework employs a mixed-initiative design, where humans and computers collaborate to create levels for 2D dungeon crawler games. We apply this framework to generate levels for three different games and analyze the results based on their expressive range, evaluating linearity and lenience. The conducted experiments demonstrate that utilizing reinforcement learning in conjunction with procedural content generation and mixed-initiative enables the generation of highly diverse levels.

Preference-Learning Emitters for Mixed-Initiative Quality-Diversity Algorithms

In mixed-initiative co-creation tasks, wherein a human and a machine jointly create items, it is important to provide multiple relevant suggestions to the designer. Quality-diversity algorithms are commonly used for this purpose, as they can provide diverse suggestions that represent salient areas of the solution space, showcasing designs with high fitness and wide variety. Because generated suggestions drive the search process, it is important that they provide inspiration, but also stay aligned with the designer's intentions. Additionally, often many interactions with the system are required before the designer is content with a solution. In this work, we tackle these challenges with an interactive constrained MAP-Elites system that leverages emitters to learn the preferences of the designer and then use them in automated steps. By learning preferences, the generated designs remain aligned with the designer's intent, and by applying automatic steps, we generate more solutions per user interaction, giving a larger number of choices to the designer and thereby speeding up the search. We propose a general framework for preference-learning emitters (PLEs) and apply it to a procedural content generation task in the video game Space Engineers. We built an interactive application for our algorithm and performed a user study with players.

Managing and controlling digital role-playing game elements: A current state of affairs

Role-playing games (RPGs) are interactive gaming experiences designed and driven by the intricate development of narratives and characters. However, these games also highlight elements, such as combat, exploration, creativity, and crafting. Game developers helped to shape the market by incorporating RPGs into digital media; however, they suffered restrictions in their game design and storytelling caused by technological limitations. Fortunately, academia has studied these challenges and mitigated these latter issues to better understand and connect digital RPGs with their analog versions. Our research offers a comprehensive overview of studies centered on technologies for managing, generating, and controlling digital RPG elements. We aim to characterize Artificial Intelligence research within this domain and elucidate the diverse techniques employed. In this context, we examined 72 papers about Procedural Content Generation (PCG) in RPG, which were identified by mixing a database search with a snowballing method. In this paper, we provide an overview of PCG techniques investigated within the realm of RPGs and present the trending approaches for future developments in the field.

Перспективы применения нейронных сетей для процедурной генерации игрового контента в киберспорте

This article provides an overview of research and practices on using neural network technologies in the field of procedural content generation for esports games. The authors explore the history, current approaches to using artificial neural networks in creating game elements, including levels, characters and scenarios, with the aim of enhancing the gaming experience and increasing the uniqueness of the virtual game world. The article covers various aspects of using neural network methods, such as generative models, autoencoders, generative adversarial networks, deep learning, and recurrent neural networks, for creating dynamic and unpredictable content. Examples of the successful implementation of these technologies in popular esports games are discussed, and potential challenges and issues related to applying neural networks in this context are explored. The authors discuss the prospects for the further development of neural network technologies in esports and offer recommendations for their optimal implementation. Overall, the article presents an analysis of the current state and future opportunities for using neural network approaches for procedural content generation in esports scenarios.

Procedural Level Generation with Diffusion Models from a Single Example

Level generation is a central focus of Procedural Content Generation (PCG), yet deep learning-based approaches are limited by scarce training data, i.e., human-designed levels. Despite being a dominant framework, Generative Adversarial Networks (GANs) exhibit a substantial quality gap between generated and human-authored levels, alongside rising training costs, particularly with increasing token complexity. In this paper, we introduce a diffusion-based generative model that learns from just one example. Our approach involves two core components: 1) an efficient yet expressive level representation, and 2) a latent denoising network with constrained receptive fields. To start with, our method utilizes token semantic labels, similar to word embeddings, to provide dense representations. This strategy not only surpasses one-hot encoding in representing larger game levels but also improves stability and accelerates convergence in latent diffusion. In addition, we adapt the denoising network architecture to confine the receptive field to localized patches of the data, aiming to facilitate single-example learning. Extensive experiments demonstrate that our model is capable of generating stylistically congruent samples of arbitrary sizes compared to manually designed levels. It suits a wide range of level structures with fewer artifacts than GAN-based approaches. The source code is available at https://github.com/shiqi-dai/diffusioncraft.

Procedural Content Generation via Knowledge Transformation (PCG-KT)

We introduce the concept of Procedural Content Generation via Knowledge Transformation (PCG-KT), a new lens and framework for characterizing PCG methods and approaches in which content generation is enabled by the process of knowledge transformation—transforming knowledge derived from one domain in order to apply it in another. Our work is motivated by a substantial number of recent PCG works that focus on generating novel content via repurposing derived knowledge. Such works have involved, for example, performing transfer learning on models trained on one game's content to adapt to another game's content, as well as recombining different generative distributions to blend the content of two or more games. Such approaches arose in part due to limitations in PCG via Machine Learning (PCGML) such as producing generative models for games lacking training data and generating content for entirely new games. In this paper, we categorize such approaches under this new lens of PCG-KT by offering a definition and framework for describing such methods and surveying existing works using this framework. Finally, we conclude by highlighting open problems and directions for future research in this area.

JAHAN: A framework for procedural generation of game maps from design specifications

Procedural content generation is a field that focuses on creating game assets employing algorithmic procedures. A critical asset in many game genres is the map. Due to the sophistication of the game map development, procedural map generation has been the subject of many studies. Considering the fundamental complexity of the game maps, many of these studies try to engage human designers in the generation process to improve the output of algorithmic procedures. However, these studies have not paid adequate attention to important non-functional requirements such as generality, comprehensiveness, controllability, and adaptability to iterative development.This paper presents a procedural map generation framework that places map designers at the center of a pipeline architecture. This framework allows designers to configure different pipeline modules and specify the characteristics of their desired map. Moreover, this framework has assertion checking features that facilitate map evaluation.We demonstrate the capabilities of this framework through one running example, along with three case studies that generate maps for first-person shooter, real-time strategy, and dungeon crawler games. Furthermore, we present a user study to show the suitability of this framework in iterative development processes and demonstrate how it reduces the labor of human designers.

Developing a framework for heterotopias as discursive playgrounds: a comparative analysis of non-immersive and immersive technologies

The discursive space represents the reordering of knowledge gained through accumulation. In the digital age, multimedia has become the language of information, and the space for archival practices is provided by non-immersive technologies, resulting in the disappearance of several layers from discursive activities. Heterotopias are unique, multilayered epistemic contexts that connect other systems through the exchange of information. This paper describes a process to create a framework for Virtual Reality, Mixed Reality, and personal computer environments based on heterotopias to provide absent layers. This study provides virtual museum space as an informational terrain that contains a “world within worlds” and presents place production as a layer of heterotopia and the subject of discourse. Automation for the individual multimedia content is provided via various sorting and grouping algorithms, and procedural content generation algorithms such as Binary Space Partitioning, Cellular Automata, Growth Algorithm, and Procedural Room Generation. Versions of the framework were comparatively evaluated through a user study involving 30 participants, considering factors such as usability, technology acceptance, and presence. The results of the study show that the framework can serve diverse contexts to construct multilayered digital habitats and is flexible for integration into professional and daily life practices.

Affective computing methods for simulation of action scenarios in video games

Video games are becoming a part of the lives of many, many people. They are able to evoke a wide range of emotions, from joy and excitement to fear and empathy, creating deep immersion and memorable impressions for players. How do they manage it and is it possible to use the emotions themselves in the game itself?This article presents the concept of developing a system of dynamic procedural content generation (PCG) with game mechanics of interaction using emotions based on methods affective computing. To achieve this goal, The article developed an emotion recognition system based on fuzzy logic. And PCG performed by Artificial Neural Networks.In the video game industry, affective computing methods, such as interaction with the expression of emotions, can bring great benefits, since the emotional involvement of the player is very important. This will make a new contribution to the diversity of human-computer interaction.

AI-Powered Procedural Content Generation: Enhancing NPC Behaviour for an Immersive Gaming Experience

AI-Powered Procedural Content Generation: Enhancing NPC Behaviour for an Immersive Gaming Experience