RTS games as test-bed for real-time AI research

2011

Video games are complex simulation environments with many real-world properties that need to be addressed in order to build robust intelligence. In particular, realtime strategy games provide a multi-scale challenge which requires both deliberative and reactive reasoning processes. Experts approach this task by studying a corpus of games, building models for anticipating opponent actions, and practicing within the game environment. We motivate the need for integrating heterogeneous approaches by enumerating a range of competencies involved in gameplay and discuss how they are being implemented in EISBot, a reactive planning agent that we have applied to the task of playing real-time strategy games at the same granularity as humans.

Proceedings of the Winter Simulation Conference 2014, 2014

Real Time Strategy (RTS) games provide complex domain to test the latest artificial intelligence (AI) research. In much of the literature, AI systems have been limited to playing one game. Although, this specialization has resulted in stronger AI gaming systems it does not address the key concerns of AI research, which focuses on the development of AI agents that can autonomously interpret, learn, and apply new knowledge. To achieve human level performance, current AI systems rely on game specific knowledge of an expert. This paper proposes a RTS language in hopes of shifting the current research focus to the development of general RTS agents. General RTS agents are AI gaming systems that can play any RTS game, defined in the proposed RTS language. The structure of the RTS language prevents game specific knowledge from being hard coded into the system, thereby facilitating research that addresses the fundamental concerns of artificial intelligence.

2013 IEEE Symposium on Foundations of Computational Intelligence (FOCI), 2013

Real-time strategy games offer a wide variety of fundamental AI research challenges. Most of these challenges have applications outside the game domain. This paper provides a review on computational intelligence in real-time strategy games (RTS). It starts with challenges in real-time strategy games, then it reviews different tasks to overcome this challenges. Later, it describes the techniques used to solve this challenges and it makes a relationship between techniques and tasks. Finally, it presents a set of different frameworks used as test-beds for the techniques employed. This paper is intended to be a starting point for future researchers on this topic.

2016

102 AI MAGAZINE StarCraft by Blizzard Entertainment is one of the most popular and famous real-time strategy (RTS) games. It provides a dynamic environment in which several agents interact to build military units with which to fight against an opponent. This game involves three distinct “races” (Protoss, Terran, and Zerg) who build different types of units and buildings and who have particular disadvantages and strengths. Players require a great deal of economic and military power to defeat their opponents while surviving and incurring minimal damage. RTS games differ from traditional board games in that they involve simultaneous movement in real time within partially observable and nondeterministic complex environments. Since the introduction of BroodWar API, StarCraft has been an important AI research platform for the development of game-playing bots using various approaches to handle a number of units and buildings by employing careful resource management and high-level tactics. ...

A commonly used technique for managing AI complexity in real-time strategy (RTS) games is to use action and/or state abstractions. High-level abstractions can often lead to good strategic decision making, but tactical decision quality may suffer due to lost details. A competing method is to sample the search space which often leads to good tactical performance in simple scenarios, but poor high-level planning. We propose to use a deep convolutional neural network (CNN) to select among a limited set of abstract action choices, and to utilize the remaining computation time for game tree search to improve low level tactics. The CNN is trained by supervised learning on game states labelled by Puppet Search, a strategic search algorithm that uses action abstractions. The network is then used to select a script — an abstract action — to produce low level actions for all units. Subsequently, the game tree search algorithm improves the tactical actions of a subset of units using a limited view of the game state only considering units close to opponent units. Experiments in the µRTS game show that the combined algorithm results in higher win-rates than either of its two independent components and other state-of-the-art µRTS agents. To the best of our knowledge, this is the first successful application of a convolutional network to play a full RTS game on standard game maps, as previous work has focused on sub-problems, such as combat, or on very small maps.

2007

We developed knowledge-rich agents to play real-time strategy games by interfacing the ORTS game engine to the Soar cognitive architecture. The middleware we developed supports grouping, attention, coordinated path finding, and FSM control of low-level unit behaviors. The middleware attempts to provide information humans use to reason about RTS games, and facilitates creating agent behaviors in Soar. Agents implemented with this system won two out of three categories in the AIIDE 2006 ORTS competition.

This paper presents a competitive combat strategy and tactics in RTS Games AI. To put it simply, if a player is building up base, he is losing out on creating an army and If he is building up his army, he is losing out on having a strong base. The key to winning, in StarCraft or any other RTS game is to balance strategy, tactics, macro and micro. To improve the game, one has to be able to keep track of everything that's going on over the entire map. And one must be able to give orders quickly and efficiently so in this paper we propose a competitive battle strategy with the help of a plot and decision tree. We simulate the strategy in MicroRTS developed in java EE by conducting a game-play between human player and MicroRTS AI (Game AI), though our proposed strategy out-performs the Game AI rarely as we did not account game playing-speed that makes a huge difference in victory but at least we succeeded in introducing a strategy that could well compete the Game AI and may defeat it but rarely.

In order to experiment with machine learning and data mining techniques in the domain of Real-Time Strategy games such as StarCraft, a dataset is required that captures the complex detail of the interactions taking place between the players and the game. This paper describes a new extraction process by which game data is extracted both directly from game log (replay) files, and indirectly through simulating the replays within the StarCraft game engine. Data is then stored in a compact, hierarchical, and easily accessible format. This process is applied to a collection of expert replays, creating a new standardised dataset. The data recorded is enough for almost the complete game state to be reconstructed, from either player's viewpoint, at any point in time (to the nearest second). This process has revealed issues in some of the source replay files, as well as discrepancies in prior datasets. Where practical, these errors have been removed in order to produce a higher-quality reusable dataset. 1 Introduction Games are an ideal domain for exploring the capabilities of Artificial Intelligence (AI) within a constrained environment and a fixed set of rules, where problem-solving techniques can be developed and evaluated before being applied to more complex real-world problems (Schaeffer 2001). Ideally, increasingly realistic games will also lead to more human-like AI being developed (Laird and van Lent 2001). Board game AI has historically received a lot of academic and public attention, but over the past decade there has been increasing interest in research based on video game AI. Real-Time Strategy (RTS) is a genre of video games in which players indirectly control many units in a simplified military simulation, which usually includes gathering resources, building infrastructure and armies, and managing units in battle. RTS games present some of the toughest challenges for AI agents, making it a difficult area for developing competent AI (Buro and Furtak 2004). It is a particularly attractive area for AI research because of how human players can quickly become adept at dealing with the complexity of the game, with experienced humans outplaying even the best academic agents (Buro and Churchill 2012).

IEEE Transactions on Games, 2018

Real-time strategy (RTS) games are a challenging application for Artificial Intelligence (AI) methods. This is because they involve simultaneous play and adversarial reasoning that is conducted in real time in large state spaces. Many AI methods for playing RTS games rely on hard-coded strategies designed by human experts. The drawback of using such strategies is that they are often unable to adapt to new scenarios during gameplay. The contribution of this paper is a new approach, called Strategy Creation via Voting (SCV), that uses a voting method to generate a large set of novel strategies from existing expert-based ones. Then, SCV uses an opponent modeling scheme during the game to choose which strategy from the generated pool of possibilities to use. By repeatedly choosing which strategy to use, SCV is able to adapt to different scenarios that might arise during the game. We implemented SCV as a bot for µRTS, a recognized RTS testbed. The results of a detailed empirical study show that SCV outperforms all approaches tested in matches played on large maps and is competitive in matches played on smaller maps.