Automatic algorithm configurators are important practical tools for improving program performance measures, such as solution time or prediction accuracy. Local search approaches in particular have proven very effective for tuning algorithms. In sequential local search, the use of predictive models has proven beneficial for obtaining good tuning results. We study the use of non-parametric models in the context of population-based algorithm configurators. We introduce a new model designed specifically for the task of predicting high-performance regions in the parameter space. Moreover, we introduce the ideas of genetic engineering of offspring as well as sexual selection of parents. Numerical results show that model-based genetic algorithms significantly improve our ability to effectively configure algorithms automatically.

Most models of Stackelberg security games assume that the attacker only knows the defender's mixed strategy, but is not able to observe (even partially) the instantiated pure strategy. Such partial observation of the deployed pure strategy -- an issue we refer to as information leakage -- is a significant concern in practical applications. While previous research on patrolling games has considered the attacker's real-time surveillance, our settings, therefore models and techniques, are fundamentally different. More specifically, after describing the information leakage model, we start with an LP formulation to compute the defender's optimal strategy in the presence of leakage. Perhaps surprisingly, we show that a key subproblem to solve this LP (more precisely, the defender oracle) is NP-hard even for the simplest of security game models. We then approach the problem from three possible directions: efficient algorithms for restricted cases, approximation algorithms, and heuristic algorithms for sampling that improves upon the status quo. Our experiments confirm the necessity of handling information leakage and the advantage of our algorithms.

Shubik's (all-pay) dollar auction is a simple yet powerful auction model that aims to shed light on the motives and dynamics of conflict escalation. Common intuition and experimental results suggest that the dollar auction is a trap, inducing conflict by its very design. However, O'Neill proved the surprising fact that, contrary to the experimental results and the intuition, the dollar auction has an immediate solution in pure strategies, i.e., theoretically it should not lead to conflict escalation. In this paper, inspired by the recent literature on spiteful bidders, we ask whether the escalation in the dollar auction can be induced by meanness. Our results confirm this conjecture in various scenarios.

We study the problem of designing envy-free sponsored search auctions, where bidders are budget-constrained. Our primary goal is to design auctions that maximize social welfare and revenue — two classical objectives in auction theory. For this purpose, we characterize envy-freeness with budgets by proving several elementary properties including consistency, monotonicity and transitivity. Based on this characterization, we come up with an envy-free auction, that is both social-optimal and bidder-optimal for a wide class of bidder types. More generally, for all bidder types, we provide two polynomial time approximation schemes (PTASs) for maximizing social welfare or revenue, where the notion of envy-freeness has been relaxed slightly. Finally, in cases where randomization is allowed in designing auctions, we devise similar PTASs for social welfare or revenue maximization problems.

Cepheus is the first computer program to essentially solve a game of imperfect information that is played competitively by humans. The game it plays is heads-up limit Texas hold'em poker, a game with over 10^14 information sets, and a challenge problem for artificial intelligence for over 10 years. Cepheus was trained using a new variant of Counterfactual Regret Minimization (CFR), called CFR+, using 4800 CPUs running for 68 days. In this paper we describe in detail the engineering details required to make this computation a reality. We also prove the theoretical soundness of CFR+ and its component algorithm, regret-matching+. We further give a hint towards understanding the success of CFR+ by proving a tracking regret bound for this new regret matching algorithm. We present results showing the role of the algorithmic components and the engineering choices to the success of CFR+.

Weighted voting games allow for studying the distribution of power between agents in situations of collective decision making. While the conventional version of these games assumes that any agent is always ready to cooperate with all others, recently, more involved models have been proposed, where cooperation is subject to restrictions. Following Myerson [1977], such restrictions are typically represented by a graph that expresses available communication links among agents. In this paper, we study the time complexity of computing two well-known power indices - the Shapley-Shubik index and the Banzhaf index - in the graph-restricted weighted voting games. We show that both are #P-complete and propose a dedicated dynamic-programming algorithm that runs in pseudo-polynomial time for graphs with the bounded treewidth.

Stackelberg game models of security have received much attention, with a number of approaches for computing Stackelberg equilibria in games with a single defender protecting a collection of targets. In contrast, multi-defender security games have received significantly less attention, particularly when each defender protects more than a single target. We fill this gap by considering a multidefender security game, with a focus on theoretical characterizations of equilibria and the price of anarchy. We present the analysis of three models of increasing generality, two in which each defender protects multiple targets. In all models, we find that the defenders often have the incentive to overprotect the targets, at times significantly. Additionally, in the simpler models, we find that the price of anarchy is unbounded, linearly increasing both in the number of defenders and the number of targets per defender. Surprisingly, when we consider a more general model, this results obtains only in a “corner” case in the space of parameters; in most cases, however, the price of anarchy converges to a constant when the number of defenders increases.

Motivated by a radically new peer review system that the National Science Foundation recently experimented with, we study peer review systems in which proposals are reviewed by PIs who have submitted proposals themselves. An (m,k)-selection mechanism asks each PI to review m proposals, and uses these reviews to select (at most) k proposals. We are interested in impartial mechanisms, which guarantee that the ratings given by a PI to others' proposals do not affect the likelihood of the PI's own proposal being selected. We design an impartial mechanism that selects a k-subset of proposals that is nearly as highly rated as the one selected by the non-impartial (abstract version of) the NSF pilot mechanism, even when the latter mechanism has the "unfair" advantage of eliciting honest reviews.

Limited lookahead has been studied for decades in perfect-information games. This paper initiates a new direction via two simultaneous deviation points: generalization to imperfect-information games and a game-theoretic approach. The question of how one should act when facing an opponent whose lookahead is limited is studied along multiple axes: lookahead depth, whether the opponent(s), too, have imperfect information, and how they break ties. We characterize the hardness of finding a Nash equilibrium or an optimal commitment strategy for either player, showing that in some of these variations the problem can be solved in polynomial time while in others it is PPAD-hard or NP-hard. We proceed to design algorithms for computing optimal commitment strategies for when the opponent breaks ties 1) favorably, 2) according to a fixed rule, or 3) adversarially. The impact of limited lookahead is then investigated experimentally. The limited-lookahead player often obtains the value of the game if she knows the expected values of nodes in the game tree for some equilibrium, but we prove this is not sufficient in general. Finally, we study the impact of noise in those estimates and different lookahead depths. This uncovers a lookahead pathology.

We study the tournament fixing problem (TFP), which asks whether a tournament organizer can rig a single-elimination (SE) tournament such that their favorite player wins, simply by adjusting the initial seeding. Prior results give two perspectives of TFP: on the one hand, deciding whether an arbitrary player can win any SE tournament is known to be NP-complete; on the other hand, there are a number of known conditions, under which a player is guaranteed to win some SE tournament. We extend and connect both these lines of work. We show that for a number of structured variants of the problem, where our player is seemingly strong, deciding whether the player can win any tournament is still NP-complete. Dual to this hardness result, we characterize a new set of sufficient conditions for a player to win a tournament. Further, we give an improved exact algorithm for deciding whether a player can win a tournament.