In multi-agent planning, agents jointly compute a plan that achieves mutual goals, keeping certain information private to the individual agents. Agents' coordination is achieved through the transmission of messages, but they can be a source of privacy leakage as they can permit a malicious agent to collect information about other agents' search processes and states. In this paper, we investigate the usage of novelty techniques in the context of (decentralised) multi-agent privacy preserving planning, addressing the challenges related to the agents' privacy and performance. In particular, we show that novelty based techniques allow a significant reduction on the number of messages transmitted among agents, increasing their privacy levels and also their performances. An experimental study analyses the effectiveness of our techniques and compares them with the state of-the-art. Finally, we examine the robustness of our approach considering different delays in the messages transmission as would occur in overloaded networks, due for example to massive attacks or critical situations.

One of the classical approaches to automated planning is the reduction to propositional satisfiability (SAT). Recently, it has been shown that incremental SAT solving can increase the capabilities of several modern encodings for SAT-based planning. In this paper, we present a further improvement to SAT-based planning by introducing a new algorithm named PASAR based on the principles of counterexample guided abstraction refinement (CEGAR). As an abstraction of the original problem, we use a simplified encoding where interference between actions is generally allowed. Abstract plans are converted into actual plans where possible or otherwise used as a counterexample to refine the abstraction. Using benchmark domains from recent International Planning Competitions, we compare our approach to different state-of-the-art planners and find that, in particular, combining PASAR with forward state-space search techniques leads to promising results.

We propose an exact algorithm for solving the longest path problem between two given vertices in undirected weighted graphs. By using graph partitioning and dynamic programming, we obtain an algorithm that is significantly faster than other state-of-the-art methods. This enables us to solve instances that were previously unsolved and solve hard instances significantly faster. We also present a parallel version of the algorithm.

After impressive successes in deterministic and fully-observable board games to significantly outclass humans, game playing research shifts towards non-deterministic and imperfect information card games, where humans are still persistently better. In this paper we devise a player that challenges human supremacy in Skat. We provide a complete player for playing selected variants of the game, with effective solutions for bidding and Skat putting, extracting knowledge from several million games. For trick play we combine expert rules with engineered tree exploration for optimal open card play. For dealing with uncertainty especially in Ouvert games we search the belief space.

A pattern database (PDB) is a pre-computed lookup table storing shortest distances from abstract states to abstract goal states. PDBs are key components in heuristic search as their entries are used to prune paths that cannot lead to an optimal solution. With the sliding-tile puzzle as an exemplary application domain, we present methods to improve the precision and size of PDBs by improving additive pattern databases to zero-aware additive pattern databases (ZPDBs), reducing the compression rate from previously 1.6 bit to 1 bit per entry, generating optimal additive pattern partitionings, and building effective collections of pattern databases. With these enhancements, we achieve an overall 8.59-fold performance gain on the 24-puzzle compared to the previously best set of 6-tile PDBs.

This paper proposes a method for learning and utilizing potentially useful interaction patterns among neighborhood-based heuristics. It is built upon a previously proposed framework designed for facilitating the task of combining multiple neighborhood-based heuristics. Basically, an algorithm derived from this framework will operate by chaining the heuristics in a pipelined fashion. Conceptually, we can view this framework as an algorithmic template that contains two user-defined components: 1) the policy H for selecting heuristics, and 2) the policy L for choosing the length of the pipeline that chains the selected heuristics. In this paper, we will develop a method that automatically derives a policy H by analyzing the experience collected from running a baseline algorithm. This analysis will distill potentially useful patterns of interactions among heuristics, and give an estimate for the frequency of using each pattern. The empirical results on three problem domains show the effectiveness of the proposed approach.

Suboptimal search algorithms can often solve much larger problems than optimal search algorithms, and thus have broad practical use. This paper returns to early algorithms like WA*, A*_e and Optimistic search. It studies the commonalities between these approaches in order to build a new bounded-suboptimal algorithm. Combined with recent research on avoiding node re-expansions in bounded-optimal search, a new solution quality bound is developed, which often provides proof of the solution bound much earlier during the search. Put together, these ideas provide a new state-of-the-art in bounded-optimal search.

Multi-agent path finding (MAPF) deals with the problem of finding collision-free paths for a set of agents. Each agent moves from its start location to its destination location in a shared environment represented by a graph. Reduction-based solving approaches for MAPF, for example reduction to SAT, exploit a time-expended layered graph, where each layer corresponds to specific time. Hence, these approaches are natural for minimizing makespan (the shortest time till all agents reach their destinations). Modeling the other frequently used objective, namely Sum of Costs (SOC; sum of paths lengths of all agents) is more difficult as the solution with the smallest SOC may not be reached in the time-expended graph with the smallest makespan. In this paper we suggest two novel approaches to estimate the makespan, that guarantees existence of a SOC-optimal solution. The approaches are empirically compared with an existing reduction-based method as well as with the state-of-the-art search-based optimal MAPF solver.

We are interested in the problem of providing intuitive instructions for human agents to enable reliable navigation in unknown environments. Since the advent of GPS and digital maps, a common approach is to visually provide a planned path on a digital map defined in terms of actions to take at specific junctions. However, this approach relies on the agent to constantly and accurately localize itself. Furthermore, it comes in stark contrast to the way humans provide instructions—by leveraging known landmarks in the environment to both augment the description of the planned path as well as to allow to detect when the agent deviated from the planned path. Hence, there is need for assurable means of localization, an intuitive way of compactly conveying directions to agents and a systematic approach to account for human errors. To this end, our key insight is to employ known landmarks in the environment to overcome these challenges. We formally model this intuitive way to use landmarks for conveying instructions and for creating contingency plans. We present experiments demonstrating the efficacy of our approach both on synthetic environments as well as on realworld maps, computed using a smart-phone iOS application that we developed.

In multi-task reinforcement learning there are two main challenges: at training time, the ability to learn different policies with a single model; at test time, inferring which of those policies applying without an external signal. In the case of continual reinforcement learning a third challenge arises: learning tasks sequentially without forgetting the previous ones. In this paper, we tackle these challenges by proposing DisCoRL, an approach combining state representation learning and policy distillation. We experiment on a sequence of three simulated 2D navigation tasks with a 3 wheel omni-directional robot. Moreover, we tested our approach's robustness by transferring the final policy into a real life setting. The policy can solve all tasks and automatically infer which one to run.