We analyze symmetric protocols to rationally coordinate on an asymmetric, efficient allocation in an infinitely repeated N-agent, C-resource allocation problems. (Bhaskar 2000) proposed one way to achieve this in 2-agent, 1-resource allocation games: Agents start by symmetrically randomizing their actions, and as soon as they each choose different actions, they start to follow a potentially asymmetric "convention" that prescribes their actions from then on. We extend the concept of convention to the general case of infinitely repeated resource allocation games with N agents and C resources. We show that for any convention, there exists a symmetric subgame perfect equilibrium which implements it. We present two conventions: bourgeois, where agents stick to the first allocation; and market, where agents pay for the use of resources, and observe a global coordination signal which allows them to alternate between different allocations. We define price of anonymity of a convention as the ratio between the maximum social payoff of any (asymmetric) strategy profile and the expected social payoff of the convention. We show that while the price of anonymity of the bourgeois convention is infinite, the market convention decreases this price by reducing the conflict between the agents.

Equilibrium computation with continuous games is currently a challenging open task in artificial intelligence. In this paper, we design an iterative algorithm that finds an ε-approximate Markov perfect equilibrium with two-player zero-sum continuous stochastic games with switching controller. When the game is polynomial (i.e., utility and state transitions are polynomial functions), our algorithm converges to ε = 0 by exploiting semidefinite programming. When the game is not polynomial, the algorithm exploits polynomial approximations and converges to an ε value whose upper bound is a function of the maximum approximation error with infinity norm. To our knowledge, this is the first algorithm for equilibrium approximation with arbitrary utility and transition functions providing theoretical guarantees. The algorithm is also empirically evaluated.

Prior work done in learning by instruction (Huffman and Laird, 1995) Learning by interacting with humans is a powerful learning demonstrated learning systems that focus on agent-initiated paradigm. In a complex world learning through self-directed interaction, where instruction is directed by impasses arising experience alone can be slow, requiring repeated interactions in a Soar agent. They noted that instructor-initiated interaction with the environment. Learning from human-agent interaction is difficult to support because of the likely interruption can reduce the complexity of the learning task by reducing of agent's reasoning.

The multiagent systems community has adopted game theory as a framework for the design of systems of multiple self-interested agents. For this to be effective, efficient algorithms must be designed to compute the solutions that game theory prescribes. In this paper, I summarize some of the state of the art on this topic, focusing particularly on how this line of work has contributed to several highly visible deployed security applications, developed at the University of Southern California.

In many real-world auctions, a bidder does not know her exact value for an item, but can perform a costly deliberation to reduce her uncertainty. Relatively little is known about such deliberative environments, which are fundamentally different from classical auction environments. In this paper, we propose a new approach that allows us to leverage classical revenue-maximization results in deliberative environments. In particular, we use Myerson (1981) to construct the first non-trivial (i.e., dependent on deliberation costs) upper bound on revenue in deliberative auctions. This bound allows us to apply existing results in the classical environment to a deliberative environment. In addition, we show that in many deliberative environments the only optimal dominant-strategy mechanisms take the form of sequential posted-price auctions.

We study a two-player pursuit-evasion game, in which an agent moving amongst obstacles is to be maintained within ``sight" of a pursuing robot. Using a discretization of the environment, our main contribution is to design an efficient algorithm that decides, given initial positions of both pursuer and evader, if the evader can take any moving strategy to go out of sight of the pursuer at any time instant. If that happens, we say that the evader wins the game. We analyze the algorithm, present several optimizations and show results for different environments. For situations where the evader cannot win, we compute, in addition, a pursuit strategy that keeps the evader within sight, for every strategy the evader can take. Finally, if it is determined that the evader wins, we compute its optimal escape trajectory and the corresponding optimal pursuit trajectory.

This paper addresses the problem of automated advice provision in settings that involve repeated interactions between people and computer agents. This problem arises in many real world applications such as route selection systems and office assistants. To succeed in such settings agents must reason about how their actions in the present influence people's future actions. The paper describes several possible models of human behavior that were inspired by behavioral economic theories of people's play in repeated interactions. These models were incorporated into several agent designs to repeatedly generate offers to people playing the game. These agents were evaluated in extensive empirical investigations including hundreds of subjects that interacted with computers in different choice selections processes. The results revealed that an agent that combined a hyperbolic discounting model of human behavior with a social utility function was able to outperform alternative agent designs. We show that this approach was able to generalize to new people as well as choice selection processes that were not used for training. Our results demonstrate that combining computational approaches with behavioral economics models of people in repeated interactions facilitates the design of advice provision strategies for a large class of real-world settings.

Decentralized partially observable Markov decision processes (Dec-POMDPs) offer a formal model for planning in cooperative multiagent systems where agents operate with noisy sensors and actuators, and local information. Prevalent Dec-POMDP solution techniques have mostly been centralized and have assumed knowledge of the model. In real world scenarios, however, solving centrally may not be an option and model parameters maybe unknown. To address this, we propose a distributed, model-free algorithm for learning Dec-POMDP policies, in which agents take turns learning, with each agent not currently learning following a static policy. For agents that have not yet learned a policy, this static policy must be initialized. We propose a principled method for learning such initial policies through interaction with the environment. We show that by using such informed initial policies, our alternate learning algorithm can find near-optimal policies for two benchmark problems.

This paper presents a theoretical advance by which factored POSGs can be decomposed into local models. We formalize the interface between such local models as the influence agents can exert on one another; and we prove that this interface is sufficient for decoupling them. The resulting influence-based abstraction substantially generalizes previous work on exploiting weakly-coupled agent interaction structures. Therein lie several important contributions. First, our general formulation sheds new light on the theoretical relationships among previous approaches, and promotes future empirical comparisons that could come by extending them beyond the more specific problem contexts for which they were developed. More importantly, the influence-based approaches that we generalize have shown promising improvements in the scalability of planning for more restrictive models. Thus, our theoretical result here serves as the foundation for practical algorithms that we anticipate will bring similar improvements to more general planning contexts, and also into other domains such as approximate planning, decision-making in adversarial domains, and online learning.

We consider symbolic dynamic programming (SDP) for solving Markov Decision Processes (MDP) with factored state and action spaces, where both states and actions are described by sets of discrete variables. Prior work on SDP has considered only the case of factored states and ignored structure in the action space, causing them to scale poorly in terms of the number of action variables. Our main contribution is to present the first SDP-based planning algorithm for leveraging both state and action space structure in order to compute compactly represented value functions and policies. Since our new algorithm can potentially require more space than when action structure is ignored, our second contribution is to describe an approach for smoothly trading-off space versus time via recursive conditioning. Finally, our third contribution is to introduce a novel SDP approximation that often significantly reduces planning time with little loss in quality by exploiting action structure in weakly coupled MDPs. We present empirical results in three domains with factored action spaces that show that our algorithms scale much better with the number of action variables as compared to state-of-the-art SDP algorithms.