Red-black planning is a recent approach to partial delete relaxation, where red variables take the relaxed semantics (accumulating their values), while black variables take the regular semantics. Practical heuristic functions can be generated from tractable sub-classes of red-black planning. Prior work has identified such sub-classes based on the black causal graph, i.e., the projection of the causal graph onto the black variables. Here, we consider cross-dependencies between black and red variables instead. We show that, if no red variable relies on black preconditions, then red-black plan generation is tractable in the size of the black state space, i.e., the product of the black variables. We employ this insight to devise a new red-black plan heuristic in which variables are painted black starting from the causal graph leaves. We evaluate this heuristic on the planning competition benchmarks. Compared to a standard delete relaxation heuristic, while the increased runtime overhead often is detrimental, in some cases the search space reduction is strong enough to result in improved performance overall.

The typical objective of path planning is to find the shortest feasible path. Many times, however, there may be no solution given the existence of constraints, such as obstacles. In these cases, the minimum constraint removal problem asks for the minimum set of constraints that need to be removed from the state space to find a solution. Unfortunately, minimum constraint removal paths do not exhibit dynamic programming properties, i.e., subsets of optimum solutions are not necessarily optimal. Thus, searching for such solutions is computationally expensive. This leads to approximate methods, which balance the cost of computing a solution and its quality. This work investigates alternatives in this context and evaluates their performance in terms of such tradeoffs. Solutions that follow a bounded-length approach, i.e., searching for paths up to a certain length, seem to provide a good balance between minimizing constraints, computational cost and path length.

We introduce an approximate search algorithm for fast maximum a posteriori probability estimation in probabilistic programs, which we call Bayesian ascent Monte Carlo (BaMC). Probabilistic programs represent probabilistic models with varying number of mutually dependent finite, countable, and continuous random variables. BaMC is an anytime MAP search algorithm applicable to any combination of random variables and dependencies. We compare BaMC to other MAP estimation algorithms and show that BaMC is faster and more robust on a range of probabilistic models.

Snake in the Box (SIB) is the problem of finding the longest simple path along the edges of an n -dimensional cube, subject to certain constraints. SIB has important applications in coding theory and communications. State of the art algorithms for solving SIB apply uninformed search with symmetry breaking techniques. We formalize this problem as a search problem and propose several admissible heuristics to solve it. Using the proposed heuristics is shown to have a huge impact on the number of nodes expanded and, in some configurations, on runtime. These results encourage further research in using heuristic search to solve SIB, and to solve maximization problems more generally.

Suboptimal heuristic search algorithms such as greedy best-first search allow us to find solutions when constraints of either time, memory, or both prevent the application of optimal algorithms such as A*. Guidelines for building an effective heuristic for A* are well established in the literature, but we show that if those rules are applied for greedy best-first search, performance can actually degrade. Observing what went wrong for greedy best-first search leads us to a quantitative metric appropriate for greedy heuristics, called Goal Distance Rank Correlation (GDRC). We demonstrate that GDRC can be used to build effective heuristics for greedy best-first search automatically.

Monte-Carlo Tree Search (MCTS) algorithms estimate the value of MDP states based on rewards received by performing multiple random simulations. MCTS algorithms can use different strategies to aggregate these rewards and provide an estimation for the states’ values. The most common aggregation method is to store the mean reward of all simulations. Another common approach stores the best observed reward from each state. Both of these methods have complementary benefits and drawbacks. In this paper, we show that both of these methods are biased estimators for the real expected value of MDP states. We propose an hybrid approach that uses the best reward for states with low noise, and otherwise uses the mean. Experimental results on the Sailing MDP domain show that our method has a considerable advantage when the rewards are drawn from a noisy distribution.

Meta-reasoning can improve numerous search algorithms, but necessitates collection of statistics to be used as probability distributions, and involves restrictive meta-reasoning assumptions. The recently suggested scheme of type systems in search algorithms is used in this paper for collecting these statistics. The statistics are then used to better estimate the unknown quantity of expected regret of computing a heuristic in Rational Lazy IDA* (RLIDA*), and also facilitate a second improvement due to relaxing one of the unrealistic meta-reasoning assumptions in RLIDA*.

This paper makes two pedagogical contributions. First, we describe two macrooperators for best-first search algorithms: the collapse macro where asubtree is deleted from memory and its best frontier value is stored in itsroot, and, the restore macro (the inverse of collapse) where thesubtree is restored to its previous structure. We show that many known searchalgorithms can be easily described by using these macros. The secondcontribution is an algorithm called Iterative Linear Best-first Search (ILBFS). ILBFS is equivalent to RBFS. While RBFS uses a recursive structure,ILBFS uses the regular structure of BFS with occasionally using the collapseand restore macros. ILBFS and RBFS are identical in the nodes that they visitand have identical properties. But, I believe that ILBFS is pedagogicallysimpler to describe and understand; it could at least serve as a pedagogicaltool for RBFS.

Learning from experience can significantly improve the performance of search based planners, especially for challenging problems like high-dimensional planning. Experience Graph (E-Graph) is a recently developed framework that encodes experiences, obtained from solving instances in the past, into a single bounded-admissible heuristic, and uses it to guide the search. While the E-Graph approach was shown to be very useful for repetitive problems, it suffers from two issues. First, computing the E-Graph heuristic is time consuming as it maintains the bounded admissibility constraints. Second, a single heuristic can get stuck in a local minimum, and thereby, degrade the performance. In this work, we present an alternative approach to improving the runtime of search from experience, based on a recently developed search algorithm Multi-heuristic A* (MHA*). This framework provides an improvement over the E-Graph planner for two reasons: a) MHA* uses multiple heuristics simultaneously to explore the search space, which reduces the probability of getting stuck in a local minimum, and b) the heuristics in MHA* can be arbitrarily inadmissible, which makes it very easy to compute them. The paper describes the framework, explains how to compute these (inadmissible) heuristics through offline and online processing and presents experimental analysis on two domains, motion planning for a 6D planar arm and large sliding tile puzzles.

Methods for learning to search for structured prediction typically imitate a reference policy, with existing theoretical guarantees demonstrating low regret compared to that reference. This is unsatisfactory in many applications where the reference policy is suboptimal and the goal of learning is to improve upon it. Can learning to search work even when the reference is poor? We provide a new learning to search algorithm, LOLS, which does well relative to the reference policy, but additionally guarantees low regret compared to deviations from the learned policy: a local-optimality guarantee. Consequently, LOLS can improve upon the reference policy, unlike previous algorithms. This enables us to develop structured contextual bandits, a partial information structured prediction setting with many potential applications.