Player coordination is a key element in many multi-player real-time digital games and cooperative real-time multi-player modes are now common in many digital-game genres. Coordination is an important part of the design of these games for several reasons: coordination can change the game balance and the level of difficulty as different types and degrees of coordination can make the game easier or more difficult; coordination is an important part of ‘playing like a team’ which affects the quality of play; and coordination as a shared activity is a key to sociality that can add to the sociability of the game. Being able to exercise control over the design of these coordination requirements is an important part of developing successful games. However, it is currently difficult to understand, describe, analyze or design coordination requirements in game situations, because current frameworks and theories do not mesh with the realities of video game design. I developed a new framework (called PLATO) that can help game designers to understand, describe, design and manipulate coordination episodes. The framework deals with five atomic aspects of coordinated activity: Players, Locations, Actions, Time, and Objects. PLATO provides a vocabulary, methodology and diagram notation for describing and analyzing coordination. I demonstrate the framework’s utility by describing coordination situations from existing games, and by showing how PLATO can be used to understand and redesign coordination requirements.

The optimal allocation of the resources to the emergency locations in the event of multiple crises in an urban environment is an intricate problem, especially when the available resources are limited. In such a scenario, it is important to allocate emergency response units in a fair manner based on the criticality of the crisis events and their requests. In this research, a crisis management tool is developed which incorporates a resource allocation algorithm. The problem is formulated as a game theoretic framework in which the crisis events are modeled as the players, the emergency response centers as the resource locations with emergency units to be scheduled and the possible allocations as strategies. The pay-off is modeled as a function of the criticality of the event and the anticipated response times. The game is played assuming a specific region within a certain locality of the crisis event to derive an optimal allocation. If a solution is not feasible, the perimeter of the locality in consideration is increased and the game is repeated until convergence. Experimental results are presented to illustrate the efficacy of the proposed methodology and metrics are derived to quantify the fairness of the solution. A regression analysis has been performed to identify the statistical significance of the results.

Parmi tous les jeux de société joués par les humains au cours de l’histoire, le jeu de go était considéré comme l’un des plus difficiles à maîtriser par un programme informatique [Van Den Herik et al., 2002]; Jusqu’à ce que ce ne soit plus le cas [Silveret al., 2016]. Cette percée révolutionnaire [Müller, 2002, Van Den Herik et al., 2002] fût le fruit d’une combinaison sophistiquée de Recherche arborescente Monte-Carlo et de techniques d’apprentissage automatique pour évaluer les positions du jeu, mettant en lumière le grand potentiel de l’apprentissage automatique pour résoudre des jeux. L’apprentissage antagoniste, un cas particulier de l’optimisation multiobjective, est un outil de plus en plus utile dans l’apprentissage automatique. Par exemple, les jeux à deux joueurs et à somme nulle sont importants dans le domain des réseaux génératifs antagonistes [Goodfellow et al., 2014] ainsi que pour maîtriser des jeux comme le Go ou le Poker en s’entraînant contre lui-même [Silver et al., 2017, Brown andSandholm, 2017]. Un résultat classique de la théorie des jeux indique que les jeux convexes-concaves ont toujours un équilibre [Neumann, 1928]. Étonnamment, les praticiens en apprentissage automatique entrainent avec succès une seule paire de réseaux de neurones dont l’objectif est un problème de minimax non-convexe et non-concave alors que pour une telle fonction de gain, l’existence d’un équilibre de Nash n’est pas garantie en général. Ce travail est une tentative d'établir une solide base théorique pour l’apprentissage dans les jeux. La première contribution explore le théorème minimax pour une classe particulière de jeux non-convexes et non-concaves qui englobe les réseaux génératifs antagonistes. Cette classe correspond à un ensemble de jeux à deux joueurs et a somme nulle joués avec des réseaux de neurones. Les deuxième et troisième contributions étudient l’optimisation des problèmes minimax, et plus généralement, les inégalités variationnelles dans le cadre de l’apprentissage automatique. Bien que la méthode standard de descente de gradient ne parvienne pas à converger vers l’équilibre de Nash de jeux convexes-concaves simples, il existe des moyens d’utiliser des gradients pour obtenir des méthodes qui convergent. Nous étudierons plusieurs techniques telles que l’extrapolation, la moyenne et la quantité de mouvement à paramètre négatif. La quatrième contribution fournit une étude empirique du comportement pratique des réseaux génératifs antagonistes. Dans les deuxième et troisième contributions, nous diagnostiquons que la méthode du gradient échoue lorsque le champ de vecteur du jeu est fortement rotatif. Cependant, une telle situation peut décrire un pire des cas qui ne se produit pas dans la pratique. Nous fournissons de nouveaux outils de visualisation afin d’évaluer si nous pouvons détecter des rotations dans comportement pratique des réseaux génératifs antagonistes. / Among all the historical board games played by humans, the game of go was considered one of the most difficult to master by a computer program [Van Den Heriket al., 2002]; Until it was not [Silver et al., 2016]. This odds-breaking break-through [Müller, 2002, Van Den Herik et al., 2002] came from a sophisticated combination of Monte Carlo tree search and machine learning techniques to evaluate positions, shedding light upon the high potential of machine learning to solve games. Adversarial training, a special case of multiobjective optimization, is an increasingly useful tool in machine learning. For example, two-player zero-sum games are important for generative modeling (GANs) [Goodfellow et al., 2014] and mastering games like Go or Poker via self-play [Silver et al., 2017, Brown and Sandholm,2017]. A classic result in Game Theory states that convex-concave games always have an equilibrium [Neumann, 1928]. Surprisingly, machine learning practitioners successfully train a single pair of neural networks whose objective is a nonconvex-nonconcave minimax problem while for such a payoff function, the existence of a Nash equilibrium is not guaranteed in general. This work is an attempt to put learning in games on a firm theoretical foundation. The first contribution explores minimax theorems for a particular class of nonconvex-nonconcave games that encompasses generative adversarial networks. The proposed result is an approximate minimax theorem for two-player zero-sum games played with neural networks, including WGAN, StarCrat II, and Blotto game. Our findings rely on the fact that despite being nonconcave-nonconvex with respect to the neural networks parameters, the payoff of these games are concave-convex with respect to the actual functions (or distributions) parametrized by these neural networks. The second and third contributions study the optimization of minimax problems, and more generally, variational inequalities in the context of machine learning. While the standard gradient descent-ascent method fails to converge to the Nash equilibrium of simple convex-concave games, there exist ways to use gradients to obtain methods that converge. We investigate several techniques such as extrapolation, averaging and negative momentum. We explore these techniques experimentally by proposing a state-of-the-art (at the time of publication) optimizer for GANs called ExtraAdam. We also prove new convergence results for Extrapolation from the past, originally proposed by Popov [1980], as well as for gradient method with negative momentum. The fourth contribution provides an empirical study of the practical landscape of GANs. In the second and third contributions, we diagnose that the gradient method breaks when the game’s vector field is highly rotational. However, such a situation may describe a worst-case that does not occur in practice. We provide new visualization tools in order to exhibit rotations in practical GAN landscapes. In this contribution, we show empirically that the training of GANs exhibits significant rotations around Local Stable Stationary Points (LSSP), and we provide empirical evidence that GAN training converges to a stable stationary point, which is a saddle point for the generator loss, not a minimum, while still achieving excellent performance.