The use of the Monte Carlo playouts as an evaluation function has proved to be a viable, general technique for searching intractable game spaces. This facilitate the use of statistical techniques like Monte Carlo Tree Search (MCTS), but is also known to require significant processing overhead. We seek to improve the quality of information extracted from the Monte Carlo playout in three ways. Firstly, by nesting the evaluation function inside another evaluation function; secondly, by measuring and utilising the depth of the playout; and thirdly, by incorporating pruning strategies that eliminate unnecessary searches and avoid traps. Our experimental data, obtained on a variety of two-player games from past General Game Playing (GGP) competitions and others, demonstrate the usefulness of these techniques in a Nested Player when pitted against a standard, optimised UCT player.

Despite recent improvements in search techniques for cost-optimal classical planning, the exponential growth of the size of the search frontier in A* is unavoidable. We investigate tiebreaking strategies for A*, experimentally analyzing the performance of standard tiebreaking strategies that break ties according to the heuristic value of the nodes. We find that tiebreaking has a significant impact on search algorithm performance when there are zero-cost operators that induce large plateau regions in the search space. We develop a new framework for tiebreaking based on a depth metric which measures distance from the entrance to the plateau, and propose a new, randomized strategy which significantly outperforms standard strategies on domains with zero-cost actions.

Identifying a small number of features that can represent the data is a known problem that comes up in areas such as machine learning, knowledge representation, data mining, and numerical linear algebra. Computing an optimal solution is believed to be NP-hard, and there is extensive work on approximation algorithms. Classic approaches exploit the algebraic structure of the underlying matrix, while more recent approaches use randomization. An entirely different approach that uses the A* heuristic search algorithm to find an optimal solution was recently proposed. Not surprisingly it is limited to effectively selecting only a small number of features. We propose a similar approach related to the Weighted A* algorithm. This gives algorithms that are not guaranteed to find an optimal solution but run much faster than the A* approach, enabling effective selection of many features from large datasets. We demonstrate experimentally that these new algorithms are more accurate than the current state-of-the-art while still being practical. Furthermore, they come with an adjustable guarantee on how different their error may be from the smallest possible (optimal) error. Their accuracy can always be increased at the expense of a longer running time.

We introduce a roommate market model, in which 2n people need to be assigned to n rooms, with two people in each room. Each person has a valuation to each room, as well as a valuation to each of other people as a roommate. Each room has a rent shared by the two people living in the room, and we need to decide who live together in which room and how much each should pay. Various solution concepts on stability and envy-freeness are proposed, with their existence studied and the computational complexity of the corresponding search problems analyzed. In particular, we show that maximizing the social welfare is NP-hard, and we give a polynomial time algorithm that achieves at least 2/3 of the maximum social welfare. Finally, we demonstrate a pricing scheme that can achieve envy-freeness for each room.

We address the control of a vertically profiling float us- ing ocean-model-based predictions of future currents. While these problems are in reality continuous control problems, we solve them by searching a discrete space of future actions. Additionally, while the environment is a continuous space, the ocean model we use is a discrete cell-based model. We show that even with an imperfect model of ocean currents, planning in the ocean current model can significantly improve results for a specific problem of controlling a vertically profiling float when trading off remaining at the same location as a virtual mooring and collecting more data with more profiles.

Model-Based Diagnosis (MBD) is a principled and domain-independent way of analyzing why a system under examination is not behaving as expected. Given an abstract description (model) of the system's components and their behavior when functioning normally, MBD techniques rely on observations about the actual system behavior to reason about possible causes when there are discrepancies between the expected and observed behavior. Due to its generality, MBD has been successfully applied in a variety of application domains over the last decades. In many application domains of MBD, testing different hypotheses about the reasons for a failure can be computationally costly, e.g., because complex simulations of the system behavior have to be performed. In this work, we therefore propose different schemes of parallelizing the diagnostic reasoning process in order to better exploit the capabilities of modern multi-core computers. We propose and systematically evaluate parallelization schemes for Reiter's hitting set algorithm for finding all or a few leading minimal diagnoses using two different conflict detection techniques. Furthermore, we perform initial experiments for a basic depth-first search strategy to assess the potential of parallelization when searching for one single diagnosis. Finally, we test the effects of parallelizing "direct encodings" of the diagnosis problem in a constraint solver.

The paper focuses on finding the m best solutions to combinatorial optimization problems using best-first or depth-first branch and bound search. Specifically, we present a new algorithm m-A*, extending the well-known A* to the m-best task, and for the first time prove that all its desirable properties, including soundness, completeness and optimal efficiency, are maintained. Since best-first algorithms require extensive memory, we also extend the memory-efficient depth-first branch and bound to the m-best task. We adapt both algorithms to optimization tasks over graphical models (e.g., Weighted CSP and MPE in Bayesian networks), provide complexity analysis and an empirical evaluation. Our experiments confirm theory that the best-first approach is largely superior when memory is available, but depth-first branch and bound is more robust. We also show that our algorithms are competitive with related schemes recently developed for the m-best task.

The current approaches to model dynamism in DCOPs solve a sequence of static problems, reacting to the changes in the environment as the agents observe them. Such approaches, thus, ignore possible predictions on the environment evolution. To overcome such limitations, we introduce the Proactive Dynamic DCOP (PD-DCOP) model, a novel formalism to model dynamic DCOPs in the presence of exogenous uncertainty. In contrast to reactive approaches, PD-DCOPs are able to explicitly model the possible changes to the problem, and take such information into account proactively, when solving the dynamically changing problem.

An effective approach to solving problems involving mixed (continuous and discrete) variables and constraints, such as hybrid systems, is to decompose them into subproblems and integrate dedicated solvers geared toward those subproblems. Here, we introduce a new framework based on a tree search algorithm to solve hybrid discrete-continuous problems that incorporates: (1) a quantum annealer that samples from the configuration space for the discrete portion and provides information about the quality of the samples, and (2) a classical computer that makes use of information from the quantum annealer to prune and focus the search as well as check a continuous constraint. We consider four variants of our algorithm, each with progressively more guidance from the results provided by the quantum annealer. We empirically test our algorithm and compare the variants on a simplified Mars Lander task scheduling problem. Variants with more guidance from the quantum annealer have better performance.

In the field of General Game Playing the imperfectinformationgames present a special challenge for researchers.In general the search space is larger, and thelack of information requires a different decision makingtechnique. A simple Monte Carlo sampling using a particlefilter may serve for the simple games, but this soonfails when more complex games are played. The HyperPlaytechnique was one such ”simple” player, soonenhanced to HyperPlay-II capable of handling the mostcomplex of games. However, this technique is very resourcehungry and may not scale well for larger games.We explore the scalability of HyperPlay-II for a varietyof imperfect-information games and test some perfectinformationpruning techniques to see if they will improveefficiency.