General Game Research Articles

Maximum Principle for General Partial Information Nonzero Sum Stochastic Differential Games and Applications

Maximum Principle for General Partial Information Nonzero Sum Stochastic Differential Games and Applications

Generalized Monte-Carlo Tree Search Extensions for General Game Playing

General Game Playing (GGP) agents must be capable of playing a wide variety of games skillfully. Monte-Carlo Tree Search (MCTS) has proven an effective reasoning mechanism for this challenge, as is reflected by its popularity among designers of GGP agents. Providing GGP agents with the knowledge relevant to the game at hand in real time is, however, a challenging task. In this paper we propose two enhancements for MCTS in the context of GGP, aimed at improving the effectiveness of the simulations in real time based on in-game statistical feedback. The first extension allows early termination of lengthy and uninformative simulations while the second improves the action-selection strategy when both explored and unexplored actions are available. The methods are empirically evaluated in a state-of-the-art GGP agent and shown to yield an overall significant improvement in playing strength.

HyperPlay: A Solution to General Game Playing with Imperfect Information

General Game Playing is the design of AI systems able to understand the rules of new games and to use such descriptions to play those games effectively. Games with imperfectinformation have recently been added as a new challenge forexisting general game-playing systems. The HyperPlay technique presents a solution to this challenge by maintaining a collection of models of the true game as a foundation for reasoning, and move selection. The technique provides existing game players with a bolt-on solution to convert from perfect-information games to imperfect-information games. In this paper we describe the HyperPlay technique, show how it was adapted for use with a Monte Carlo decision making process and give experimental results for its performance.

Instability of defection in the prisoner's dilemma under best experienced payoff dynamics

We study population dynamics under which each revising agent tests each action k times, with each trial being against a newly drawn opponent, and chooses the action whose mean payoff was highest during the testing phase. When k=1, defection is globally stable in the prisoner's dilemma. By contrast, when k>1 we show that, if the gains from defection are not too large, there exists a globally stable state in which agents cooperate with probability between 28% and 50%. Next, we characterize stability of strict equilibria in general games. Our results demonstrate that the empirically plausible case of k>1 can yield qualitatively different predictions than the case k=1 commonly studied in the literature.

Research Summary

Monte-Carlo Tree Search (MCTS) is an online planning algorithm that combines the ideas of best-first tree search and Monte-Carlo evaluation. Since MCTS is based on sampling, it does not require a transition function in explicit form, but only a generative model of the domain. Because it grows a highly selective search tree guided by its samples, it can handle huge search spaces with large branching factors. By using Monte-Carlo playouts, MCTS can take long-term rewards into account even with distant horizons. Combined with multi-armed bandit algorithms to trade off exploration and exploitation, MCTS has been shown to guarantee asymptotic convergence to the optimal policy, while providing approximations when stopped at any time. The relatively new MCTS approach has started a revolution in computer Go. Furthermore, it has achieved considerable success in domains as diverse as the games of Hex, Amazons, LOA, and Ms. Pacman; in General Game Playing, planning, and optimization. Whereas the focus of previous MCTS research has been on the practical application, current research begins to address the problem of understanding the nature, the underlying principles, of MCTS. A careful understanding of MCTS will lead to more effective search algorithms. Hence, my two interrelated research questions are: How can we formulate models that increase our understanding of how MCTS works? and How can we use the developed understanding to create effective search algorithms? This research summary describes the first steps I undertook in these directions, as well as my plans for future work.

Fast and Diverse Narrative Planning through Novelty Pruning

Novelty pruning is a simple enhancement that can be added to most planners. A node is removed unless it is possible to find a set of n literals which are true in the current state and have never all been true in any of that plan's previous states. Expanding on the success of the Iterated Width algorithm in classical planning and general game playing, we apply this technique to narrative planning. Using a suite of 8 benchmark narrative planning problems, we demonstrate that novelty pruning can be used with breadth-first search to solve smaller problems optimally and combined with heuristic search to solve larger problems faster. We also demonstrate that when many solutions to the same problem are generated, novelty pruning can produce a wider variety of solutions in some domains.

Towards General RPG Playing

General videogame playing has come a long way in a short period of time, but remains at the level of solving relatively short games made up of distinct and isolated episodes. Even simple console role-playing games (RPGs) are far beyond the reach of current techniques, requiring the synthesis of cultural knowledge with compositional reasoning over several interconnected sub-games. We explore how the challenges of playing these games could spark new advances in compositional analysis of games and common-sense reasoning. General RPG playing can leverage advances in episodic general game playing and in areas like text understanding, image classification, and automated game design learning. It has direct applications in design support and AI-based game design, and the techniques used to enable it could generalize to other families of games such as adventure, open-world, and simulation games. In this paper, we describe the motivation behind general RPG playing in a sub-domain of Nintendo Entertainment System (NES) RPGs, some promising approaches to some of its fundamental issues, and immediate next steps; we conclude by describing a few concrete benchmark problems on the path towards automated play of these complex games.

An Algorithmic Approach to Decorative Content Placement

Placing decorative content in any virtual environment is not a trivial task, as the location of even non-critical content influences player movement choices through the level, as well as the general game aesthetic. In this work we use path visibility properties to define and investigate three methods for content placement in a game level, aimed at generating shared, unique, and never seen content. We compare and evaluate the three approaches using a human study where each player was given a level layout based on our three methods, showing our directed process has a non-trivial and controllable impact on player movement.

The potential of sports and play activities in relation to the education of young adolescents in the desire for independence in general secondary education institutions

The article reveals the potential of sports and play activities for the education of young adolescents' desire for independence in general secondary education. The presentation of the main material is based on the results of a preliminary theoretical study of the essence of the concept of the desire of younger adolescents to independence in the process of sports and play activities. On the basis of certain criteria of upbringing in younger adolescents the desire for independence in the process of sports and games. In the process of physical education in general secondary education games are used as a means of physical culture and as a form of organization of classes. Sports games - a system of characteristic physical exercises, which are performed in different variants according to changing game situations and are evaluated for effectiveness the end result of motor actions. An important feature of sports games are complex collective actions aimed at achieving the planned result. Most types of sports games are team games, so success in competitions largely depends on the coherence of all participants in rapidly changing game situations. Complicating their interactions is the fact that during the game there is no predetermined sequence of game techniques and a constant rhythm. Depending on the game situation, there is a constant change in the emotional state of the player, in connection with which he must be able to control their emotions, actions, mental state in various unforeseen factors.

General Deep Reinforcement Learning in NES Games

The techniques involved in general game playing with Artificial Intelligence (AI) have advanced to meet the challenges of the most popular video game and board game domains. Until recently, the video game domains used as testbeds have been relatively simple. That is, the most complex console-wide domain that has been solved using Deep Reinforcement Learning is the Atari domain, which is decades behind modern video game domains. This work explores a more complex domain (the Nintendo Entertainment System, or NES) and the associated difficulties in developing deep reinforcement learning agents for it. To understand these difficulties, we trained agents on NES games with little to no expert knowledge provided to the agents. After developing some understanding of the challenges of the domain, we suggest areas on which to focus to solve this domain, work which will hopefully lead the field to solving ever more complex environments both real world and theoretical. This paper determines some of the necessary changes in hyperparameters and reward functions to solve the NES domain compared to the more popular Atari domain while using game pixel data as the only inputs to the agents’ neural networks.

Grouping Students by Skill Level in Mini-Volleyball: Effect on Game Performance and Knowledge in Sport Education.

The purpose of this study was to explore any differences in game performance variables and knowledge among a cohort of high school students who participated in either homogeneous or heterogeneous skill level groups (N = 126) across a 12-lesson mini-volleyball sport education unit of study. This study followed a mixed-methods approach using a quasi-experimental pre-test/post-test design. The quantitative variables analyzed were decision making, skill execution, game performance, game involvement, and game knowledge. We also evaluated students' performance qualitatively, employing two methods: (a) experts' analysis of students' game performance, and (b) students' and teachers' perceptions of students' performance. We analyzed quantitative data through a series of paired samples t-tests comparing pre- and post-test scores according to the grouping strategy. Students became more competent in their game play and more knowledgeable in their technique, the sport's rules, tactical awareness, and general game knowledge. However, grouping students by skill level had no impact on gains in game performance variables and knowledge. Although sport education literature shows a preference for heterogeneity in ability-based grouping, within our data both heterogeneous and homogenous groups of higher and lower skilled students achieved improvements in game performance and knowledge, leading us to suggest that teachers who are interested in grouping students to create a meaningful learning experience should consider criteria other than student ability.

∃ℝ-complete Decision Problems about (Symmetric) Nash Equilibria in (Symmetric) Multi-player Games

A decision problem about Nash equilibria is concerned with whether a given game has a Nash equilibrium with certain natural properties. We settle the complexity of such decision problems over multi-player games, establishing that (nearly) all decision problems that were before shown NP -complete over 2-player games in References [5, 12, 18] become ∃ℝ-complete over multi-player games. ∃ℝ [27] is the class capturing the complexity of deciding the Existential Theory of the Reals . Specifically, we present a simple, unifying, parametric polynomial time reduction from the problem of deciding, given a 3-player (symmetric) game, whether there is a (symmetric) Nash equilibrium where all strategies played with non-zero probability come from a given set, which was shown ∃ℝ-complete in Reference [17]. By suitably working on the tuning parameters, our reduction delivers two extensive catalogs of ∃ℝ-complete decision problems in multi-player games. The first addresses Nash equilibria in general games, while the second encompasses symmetric Nash equilibria in symmetric games. These results resolve completely the main open problem from Reference [17] to enlarge the class of ∃ℝ-complete decision problems about (symmetric) Nash equilibria in multi-player (symmetric) games.

Exploring Monte Carlo Negotiation Search with Nontrivial Agreements

The application of automated negotiations to general game playing is a research area with far-reaching implications. Non-zero sum games can be used to model a wide variety of real-world scenarios and automated negotiation provides a framework for more realistically modeling the behavior of agents in these scenarios. A particular recent development in this space is the Monte Carlo Negotiation Search (MCNS) algorithm, which can negotiate to find valuable cooperative strategies for a wide array of games (such as those of the Game Description Language). However, MCNS only proposes agreements corresponding to individual sequences of moves without any higher-level notions of conditional or stateful strategy. Our work attempts to lift this restriction. We present two contributions: extensions to the MCNS algorithm to support more complex agreements and an agreement language for GDL games suitable for use with our algorithm. We also present the results of a preliminary experiment in which we use our algorithm to search for an optimal agreement for the iterated prisoners dilemma. We demonstrate significant improvement of our algorithm over random agreement sampling, although further work is required to more consistently produce optimal agreements.

Cross-layer tradeoff of QoS and security in Vehicular ad hoc Networks: A game theoretical approach

Vehicular ad hoc Networks (VANETs) have high requirements on both Quality of Service (QoS) and security due to their unique wireless features such as highly dynamic vehicles and unreliable channels. However, it is challenging to jointly optimize QoS and security because they are conflicting objectives that contend for limited network resources. Therefore, it is essential to make a tradeoff between them. In this study, a two-period game is designed to deal with the external decision-making of vehicles and internal resource competition of an individual vehicle, respectively. In the first-period external decision-making, each vehicle aims to maximize the throughput by constructing a cross-layer utility between Medium Access Control (MAC) and Physical (PHY) layers. Moreover, the optimal transmission power of the external decision-making is obtained as the vehicle’s best strategy, which serves as an input in the next period. In the second-period internal decision-making, the conflicting objectives, i.e., QoS and security, inside an individual vehicle are viewed as two abstract players who compete for the vehicle’s limited resources. Furthermore, a general game is formulated for the internal decision-making to model the behaviors of the “communication player” in controlling the number of data blocks and that of the “security player” in deciding the hash length. The theoretical value of the Nash equilibrium of the internal game is derived and proved to be the two players’ optimal strategy composition. The simulation results demonstrate that the proposed game enables a vehicle to obtain satisfactory QoS and security levels by dynamically adjusting its optimal strategies according to the situated network and traffic environment.

Balanced Externalities and the Proportional Allocation of Nonseparable Contributions

In this paper, we study the implications of extending the balanced cost reduction property from queueing problems to general games. As a direct translation of the balanced cost reduction property, the axiom of balanced externalities for solutions of games, requires that the payoff of any player is equal to the total externality she inflicts on the other players with her presence. We show that this axiom and efficiency together characterize the Shapley value for 2-additive games. However, extending this axiom in a straightfoward way to general games is incompatible with efficiency. Keeping as close as possible to the idea behind balanced externalities, we weaken this axiom by requiring that every player's payoff is the same fraction of its total externality inflicted on the other players. This weakening, which we call weak balanced externalities, turns out to be compatible with efficiency. More specifically, the unique efficient solution that satisfies this weaker property is the proportional allocation of nonseparable contribution (PANSC) value, which allocates the total worth proportional to the separable costs of the players. We also provide characterizations of the PANSC value using a reduced game consistency axiom.

Essential stability of the alpha cores of finite games with incomplete information

We introduce a variant of Milgrom and Weber’s (1985) model of n-person games with incomplete information (games for short) and define a correspondence that maps each game to its α-core (the α-core correspondences). Our main objective is to prove such a correspondence to be generically lower semicontinuous. For a multi-valued solution correspondence, the lower semicontinuity is relevant as a theoretical base for predicting outcomes using game-theoretic models. We introduce a family of games parametrized by both payoff functions and information structures (common priors), which allows simultaneous perturbations in those two parameters. We then appeal to Fort’s (1951) theorem to conclude that generic games are essential relative to the parameter space.

Evolutionary game analysis of green industry development in the age of sharing economy

ABSTRACT Sharing economy is also a kind of green economy in essence, which brings opportunities for the development of resource-sharing green industry. By constructing a general evolutionary game model of green industry, the evolutionary stability strategy of green industry system is obtained, and the internal mechanism of the spontaneous evolution of green industry is explained. Furthermore, by comparing the evolutionary game model and numerical simulation results of two green industry development models of factor input and resource sharing in the era of sharing economy, it is found that compared with the factor input development model, the resource-sharing development model is easier to realize the evolutionary game equilibrium between enterprises and consumers. Enriching government subsidies and strengthening consumer guide will help promote the development of resource-sharing green industry.

Physiological and technical demands of the small-sided and generic games in female futsal players

Aim: This study aimed to compare the heart rate (HR) values and technical actions frequencies during small-sided (SSG) and generic (GG) games in women futsal players. Methods: Six futsal players (age: 20.5 ± 7.4 years, height: 163.3 ± 16 cm, body mass: 57.9 ± 22 kg, maximum rate of oxygen consumption (VO2max): 42.0 ± 9 mL.kg.min−1) performed two experimental conditions in two days separated for at least 72h of interval: (1) one set of 10 min of SSG, 3x3 players; and (2) one set of 10 min of GG, 4 (1 joker) x 4 (1 joker) players. During each experimental session, the subject's HR values were monitored every 1 min of play. Pass, shot, dribbling, driving, and disarmament were recorded during the games. Results: There was no difference between SSG and GG for HR average (p = 0.50). The HR values that were recorded at the first minute of GG were higher than SSG (p = 0.02). The HRmax in SSG and GG were equivalent to 91.9% and 90.8% of HRmax of the incremental test, respectively. Total technical actions were higher in SSG (+ 18.4%, p = 0.001) than GG. Pass was higher in the SSG (+16%; p = 0.01) in comparison to GG. Conclusion: Both models were higher than HRmax equivalent to AT velocity. The technical actions of the SSG were higher than GG, suggesting that the SSG model is a suitable technical and physiological stimulus for the development of female futsal players.

ПОПУЛЯРНА КУЛЬТУРА США В ДЗЕРКАЛІ КОНВЕНЦІЙ ВІДЕОІГРОВОЇ ІНДУСТРІЇ

Video game culture is one of the most dynamically developed forms of modern popular culture. Over the past 10 years, the number of communities that represent certain areas has grown rapidly. Being in direct connection with film industry, literature and music the video game culture has occupied a new segment that requires detailed research. New opportunities of online communication have transformed the traditional understanding of gaming culture. Social segment identifies as gamers. This kind of culture can be used to spread a social message using human interactions. It should be noted that in a comprehensive sense, this problem covers the interaction with computer and algorithm art, artmedia, demoscene, cybernetic and digital art, evolutionary and generative art, new media, systems art and other areas which are represented in PC, VR, Console, Handheld, Tabletop. Communities of gamers, as well as game developers have created and promoting quite recognizable festivals and conventions which are described in this article. Games and game-related merchandise by Sony Interactive Entertainment – PlayStation Experience; BlizzCon by Blizzard Entertainment with their major franchises: Overwatch, Warcraft, StarCraft, Heroes of the Storm et al; general video gaming and video game streaming – TwitchCon convention; RTX conventions by Rooster Teeth; Penny Arcade Expo as gaming culture festivals with their PAX Arena, LAN Party, eSports tournament and Omegathon. This article does not describe the US comic con conventions that carry much more content than the focus on video games, such as San Diego Comic-Con International, New York Comic Con et al. Gamers have developed specific communication languages and rituals, they spread their culture on gaming conventions and discussion websites that create a distinct virtual space. An important part of video game culture is the activity of gamers on YouTube, gaming channels are very popular, this is confirmed by millions of subscribers. Post-dynamic structure of changes was formed in all streams of public life during the growth of the pandemic. COVID-19 is accelerating existing trends and indicate future transformations in the video game industry.

Polynomial Time Method for Solving Nash Equilibria of Zero-Sum Games

There are a few studies that focus on solution methods for finding a Nash equilibrium of zero-sum games. We discuss the use of Karmarkar’s interior point method to solve the Nash equilibrium problems of a zero-sum game, and prove that it is theoretically a polynomial time algorithm. We implement the Karmarkar method, and a preliminary computational result shows that it performs well for zero-sum games. We also mention an affine scaling method that would help us compute Nash equilibria of general zero-sum games effectively.