Proceedings of the International Symposium on Combinatorial Search

Efficient and Exact Public Transport Routing via a Transfer Connection Database

Abdallah Abuaisha — 2024-06-01

We explore the earliest arrival time problem in public transport journey planning. A journey typically consists of multiple scheduled public transport legs. The actual time required to transfer between these legs can substantially influence route planning. Therefore, we properly model transfers by incorporating their exact costs. We then introduce a novel oracle-based routing algorithm that constructs an efficient transfer database, considering the proposed transfer model. The database is leveraged online to quickly reconstruct the optimal journey in response to an earliest arrival time query. Our experimental results show that neglecting exact transfer costs often lead to either infeasible or suboptimal route plans. Furthermore, the findings highlight the efficiency of our algorithm in handling queries, demonstrated by response times within mere microseconds.

Heuristic Search for the Orienteering Problem with Time-Varying Reward

Chao Cao — 2024-06-01

The Orienteering Problem (OP) seeks a path on a graph to maximize total rewards collected subject to a path length budget. Typically, a reward is achieved by visiting a vertex in the graph, and such a reward is constant for all time. This paper considers a variant of OP where the reward of each vertex is an arbitrary time-dependent function, and hence the name time-varying reward OP (TR-OP). To solve this problem, we develop a novel heuristic search algorithm called Reward Maximization A* (RMA*), which is guaranteed to find an optimal solution to TR-OP. We also develop a fast method to compute an admissible heuristic for RMA* that can effectively direct the search to save computational effort. Furthermore, we introduce a hyper-parameter in RMA* that trades off between solution quality and runtime efficiency for RMA*. We benchmark RMA* against a recent dynamic programming (DP) approach, which runs fast in practice, but has no guarantee of the solution optimality. In our tests, RMA* reduces the runtime by up to 70% compared to DP. By adjusting the hyper-parameter, RMA* is able to find solutions with up to 30% more rewards than those found by DP.

Avoiding Node Re-Expansions Can Break Symmetry Breaking

Mark Carlson — 2024-06-01

Symmetry breaking and weighted-suboptimal search are two popular speed up techniques used in pathfinding search. It is a commonly held assumption that they are orthogonal and easily combined. In this paper we illustrate that this is not necessarily the case when combining a number of symmetry breaking methods, based on Jump Point Search, with Weighted A*, a bounded suboptimal search approach which does not require node re-expansions. Surprisingly, the combination of these two methods can cause search to fail, finding no path to a target node when clearly such paths exist. We demonstrate this phenomena and show how we can modify the combination to always succeed with low overhead.

Generalized Longest Simple Path Problems: Speeding up Search Using SPQR Trees

Gal Dahan — 2024-06-01

The longest simple path and snake-in-a-box are combinatorial search problems of considerable research interest. Recent work has recast these problems as special cases of a generalized longest simple path (GLSP) framework, and showed how to generate improved search heuristics for them. The greatest reduction in search effort was based on SPQR tree rules, but it was posed as an open problem how to use them optimally. Unrelated to search, a theoretical paper on the existence of simple cycles that include three given edges answers such queries in linear time with SPQR trees. These theoretical results are utilized in this paper to develop advanced heuristics and search partitioning for GLSP. Empirical results on grid-based graphs show that these heuristics can result in orders of magnitude reduction in the number of expansions, as well as significantly reduced overall runtime in most cases.

Introducing Delays in Multi Agent Path Finding

Justin Kottinger — 2024-06-01

We consider a Multi-Agent Path Finding (MAPF) setting where agents have been assigned a plan, but during its execution some agents are delayed. Instead of replanning from scratch when such a delay occurs, we propose delay introduction, whereby we delay some additional agents so that the remainder of the plan can be executed safely. We show that finding the minimum number of additional delays is APX-hard. However, in practice we can find optimal delay-introductions using Conflict-Based Search for very large numbers of agents, and both planning time and the resulting length of the plan are comparable, and sometimes outperform the state-of-the-art heuristics for replanning.

Solving Facility Location Problems via FastMap and Locality Sensitive Hashing

Ang Li — 2024-06-01

Facility Location Problems (FLPs) arise while serving multiple customers in a shared environment, minimizing transportation and other costs. Hence, they involve the optimal placement of facilities. They are defined on graphs as well as in Euclidean spaces with or without obstacles; and they are typically NP-hard to solve optimally. There are many heuristic algorithms tailored to different kinds of FLPs. However, FLPs defined in Euclidean spaces without obstacles are the most amenable to efficient and effective heuristic algorithms. This motivates the idea of quickly reformulating FLPs on graphs and in Euclidean spaces with obstacles to FLPs in Euclidean spaces without obstacles. Towards this end, we propose a new approach that uses FastMap and Locality Sensitive Hashing. FastMap is a near-linear-time algorithm that embeds the vertices of a graph in a Euclidean space while approximately preserving graph-based distances as Euclidean distances for all pairs of vertices. Through extensive experiments, we show that our approach significantly outperforms other state-of-the-art competing algorithms on a variety of FLPs: the Multi-Agent Meeting, Vertex K-Median (VKM), Weighted VKM, and the Capacitated VKM problems.

Modeling Assistance for Hierarchical Planning: An Approach for Correcting Hierarchical Domains with Missing Actions

Songtuan Lin — 2024-06-01

The complexity of modeling planning domains is a major obstacle for making automated planning techniques more accessible, raising the demand of tools for providing modeling assistance. In particular, tools that can automatically correct errors in a planning domain are of great importance. Previous works have devoted efforts to developing such approaches for correcting classical (non-hierarchical) domains. However, no approaches exist for hierarchical planning, which is what we offer here. More specifically, our approach takes as input a flawed hierarchical domain together with a plan known to be a solution but actually contradicting the domain (due to errors in the domain) and outputs corrections to the domain that add missing actions to the domain which turn the plan into a solution. The approach achieves this by compiling the problem of finding corrections to another hierarchical planning problem.

Multi-Agent Path Execution with Uncertainty

Yihao Liu — 2024-06-01

In real-world multi-agent applications, unexpected conditions can break the assumptions made in path planning and degrade the effectiveness of path execution. This paper studies robust and effective execution of multi-agent path plans under uncertainty. To guarantee conflict-freeness and deadlock-freeness, we define a feasibility problem to check whether the remaining portion of a path plan can be successfully executed. We prove that the problem is NP-complete and propose a feasibility test algorithm. We further develop algorithms to coordinate the agents online and have as many of them as possible moving concurrently to maximize the effectiveness of execution. We experimentally demonstrate the path execution effectiveness and computational efficiency of our algorithms.

A Deterministic Search Approach for Solving Stochastic Drone Search and Rescue Planning Without Communications

Evgeny Mishlyakov — 2024-06-01

In disaster relief efforts, delivering aid to areas with no communication poses a significant challenge. Unmanned aerial vehicles (UAVs) can be utilized to deliver aid kits to survivors in hard-to-reach areas; unfortunately, in some areas, lack of communication and infrastructure presents a key problem. In this paper, we address a stochastic planning problem of planning for a set of UAVs that deliver aid kits to areas that lack communications, where we do not know in advance the locations where aid kits need to be delivered, but rather have probabilistic information about the locations of aid targets. Our main insight is that, despite the stochastic nature of this problem, we can solve it through deterministic search by monitoring the expected reward for each partial solution. This insight enables the application of deterministic planning techniques, empirically demonstrating a notable improvement in efficiency and response speed. Our approach presents a promising solution to addressing the challenge of delivering aid in regions with limited radio infrastructure, as well as similar planning problems.

Prioritised Planning with Guarantees

Jonathan Morag — 2024-06-01

Prioritised Planning (PP) is a family of incomplete and sub-optimal algorithms for multi-agent and multi-robot navigation. In PP, agents compute collision-free paths in a fixed order, one at a time. Although fast and usually effective, PP can still fail, leaving users without explanation or recourse. In this work, we give a theoretical and empirical basis for better understanding the underlying problem solved by PP, which we call Priority Constrained MAPF (PC-MAPF). We first investigate the complexity of PC-MAPF and show that the decision problem is NP-hard. We then develop Priority Constrained Search (PCS), a new algorithm that is both complete and optimal with respect to a fixed priority ordering. We experiment with PCS in a range of settings, including comparisons with existing PP baselines, and we give first-known results for optimal PC-MAPF on a popular benchmark set.

Curriculum Generation for Learning Guiding Functions in State-Space Search Algorithms

Sumedh Pendurkar — 2024-06-01

This paper investigates methods for training parameterized functions for guiding state-space search algorithms. Existing work commonly generates data for training such guiding functions by solving problem instances while leveraging the current version of the guiding function. As a result, as training progresses, the guided search algorithm can solve more difficult instances that are, in turn, used to further train the guiding function. These methods assume that a set of problem instances of varied difficulty is provided. Since previous work was not designed to distinguish the instances that the search algorithm can solve from those that cannot be solved with the current guiding function, the algorithm commonly wastes time attempting and failing to solve many of these instances. In this paper, we improve upon these training methods by generating a curriculum for learning the guiding function that directly addresses this issue. Namely, we propose and evaluate a Teacher-Student Curriculum (TSC) approach where the teacher is an evolutionary strategy that attempts to generate problem instances of ``correct difficulty'' and the student is a guided search algorithm utilizing the current guiding function. The student attempts to solve the problem instances generated by the teacher. We conclude with experiments demonstrating that TSC outperforms the current state-of-the-art Bootstrap Learning method in three representative benchmark domains and three guided search algorithms, with respect to the time required to solve all instances of the test set.

Optimised Variants of Polynomial Compilation for Conditional Effects in Classical Planning

Francesco Percassi — 2024-06-01

Conditional effects are a key feature in classical planning, enabling the description of actions whose outcomes are state-dependent. It is well known that the polynomial removal of conditional effects necessarily increases the size of a valid plan by a polynomial factor while preserving exactly the plan size requires an exponential encoding of the problem. The paper proposes and empirically evaluates optimisations for existing polynomial compilations. These optimisations aim to make the resulting compilations more suitable for planners while limiting the increase in plan size, which is inevitable if we want to keep the compilation polynomial. Specifically, the paper introduces a polynomial compilation technique that expands conditional effects when their number is below a certain threshold and sequentialises them otherwise. Additionally, the paper demonstrates that even straightforward optimisations can have a notable impact.

Unconstraining Multi-Robot Manipulation: Enabling Arbitrary Constraints in ECBS with Bounded Sub-Optimality

Yorai Shaoul — 2024-06-01

Multi-Robot-Arm Motion Planning (M-RAMP) is a challenging problem featuring complex single-agent planning and multi-agent coordination. Recent advancements in extending the popular Conflict-Based Search (CBS) algorithm have made large strides in solving Multi-Agent Path Finding (MAPF) problems. However, fundamental challenges remain in applying CBS to M-RAMP. A core challenge is the existing reliance of the CBS framework on conservative "complete" constraints. These constraints ensure solution guarantees but often result in slow pruning of the search space -- causing repeated expensive single-agent planning calls. Therefore, even though it is possible to leverage domain knowledge and design incomplete M-RAMP-specific CBS constraints to more efficiently prune the search, using these constraints would render the algorithm itself incomplete. This forces practitioners to choose between efficiency and completeness. In light of these challenges, we propose a novel algorithm, Generalized ECBS, aimed at removing the burden of choice between completeness and efficiency in MAPF algorithms. Our approach enables the use of arbitrary constraints in conflict-based algorithms while preserving completeness and bounding sub-optimality. This enables practitioners to capitalize on the benefits of arbitrary constraints and opens a new space for constraint design in MAPF that has not been explored. We provide a theoretical analysis of our algorithms, propose new "incomplete" constraints, and demonstrate their effectiveness through experiments in M-RAMP.

Neural Sequence Generation with Constraints via Beam Search with Cuts: A Case Study on VRP

Pouya Shati — 2024-06-01

In recent years, neural sequence models have been applied successfully to solve combinatorial optimization problems. Solutions, encoded as sequences, are typically generated from trained models via beam search, a search algorithm that generates sequences token-by-token while keeping a fixed number of promising partial solutions at each step. In this paper, we explore the problem of augmenting beam search generation with the enforcement of requirements---hard constraints that any generated solution must adhere to. We propose a hybrid approach, by encoding the requirements in the form of a constraint satisfaction problem (CSP) and iteratively solving the CSP to cut any partial solution within the beam search that is incapable of satisfying the requirements. We study this problem in the context of vehicle routing problems (VRP) further augmented with capacity-related or temporal requirements. We experimentally show that cuts often allow us to satisfy the requirements with negligible impact on solution quality. Without the use of cuts, beam search is shown to be exponentially less likely to satisfy the requirements as the length of the solution increases and/or the requirements are strengthened.

On the Properties of All-Pair Heuristics

Shahaf Shperberg — 2024-06-01

While most work in heuristic search concentrates on goal-specific heuristics, which estimate the shortest path cost from any state to the goal, we explore all-pair heuristics that estimate distances between all pairs of states. We examine the relationship between these heuristic functions and the shortest distance function they estimate, revealing that all-pair consistent heuristics may violate the triangle inequality. Thus, we introduce a new property for heuristics called Δ-consistency, requiring adherence to the triangle inequality. Additionally, we present a method for transforming standard consistent heuristics to be Δ-consistent, showcasing its benefits through a synthetic example. We then show that common heuristic families inherently exhibit Δ-consistency. This positive finding encourages the use of all-pair consistent heuristics, and prompts further investigation into the optimality of A*, when given an all-pair heuristic instead of a goal-specific heuristic.

ITA-ECBS: A Bounded-Suboptimal Algorithm for Combined Target-Assignment and Path-Finding Problem

Yimin Tang — 2024-06-01

Multi-Agent Path Finding (MAPF), i.e., finding collision-free paths for multiple robots, plays a critical role in many applications. Sometimes, assigning a target to each agent also presents a challenge. The Combined Target-Assignment and Path-Finding (TAPF) problem, a variant of MAPF, requires one to simultaneously assign targets to agents and plan collision-free paths for agents. Several algorithms, including CBM, CBS-TA, and ITA-CBS, optimally solve the TAPF problem, with ITA-CBS being the leading algorithm for minimizing flowtime. However, the only existing bounded-suboptimal algorithm ECBS-TA is derived from CBS-TA rather than ITA-CBS. So, it faces the same issues as CBS-TA, such as searching through multiple constraint trees and spending too much time on finding the next-best target assignment. We introduce ITA-ECBS, the first bounded-suboptimal variant of ITA-CBS. Transforming ITA-CBS to its bounded-suboptimal variant is challenging because different constraint tree nodes can have different assignments of targets to agents. ITA-ECBS uses focal search to achieve efficiency and determines target assignments based on a new lower bound matrix. We show that it runs faster than ECBS-TA in 87.42% of 54,033 test cases.

The Bench Transition System and Stochastic Exploration

Dawson Tomasz — 2024-06-01

Stochastic exploration has been shown to be an effective way to mitigate the negative impact that heuristic local minima and plateaus can have on Greedy Best First Search (GBFS). Previous work has induced exploration using type systems, which typically partition the state-space using simple features like heuristic value and depth. In this work, we introduce new type systems motivated by the Bench Transition System (BTS). The BTS is a structure used to characterize the behaviour of GBFS, that is based on high water-mark benches, which are sets of states that have made the same amount of progress towards the goal. Since the BTS cannot be constructed during search, our type systems approximate the BTS using the notions of Heuristic Improvement and Low Water-Mark. We first identify that these approximations are exact in state-spaces with plateaus but no local minima, and also show that the resulting type systems are probabilistically complete. Our empirical evaluation shows the effectiveness of this approach on a variety of planning domains.

Clique Analysis and Bypassing in Continuous-Time Conflict-Based Search

Thayne T. Walker — 2024-06-01

While the study of unit-cost Multi-Agent Pathfinding (MAPF) problems has been popular, many real-world problems require continuous time and costs. In this context, this paper studies symmetry-breaking enhancements for Continuous-Time Conflict-Based Search (CCBS), a solver for continuous-time MAPF. Resolving conflict symmetries in MAPF can require an exponential amount of work. We adapt known symmetry-breaking enhancements from unit-cost domains for CCBS: bypassing and biclique constraints. We then improve upon these to produce a new state-of-the-art algorithm: CCBS with disjoint k-partite cliques (CCBS+DK). Finally, we show empirically that CCBS+DK solves for up to 20% more agents in the same amount of time when compared to previous state-of-the-art.

Real-time Safe Interval Path Planning

Devin Wild Thomas — 2024-06-01

Navigation among dynamic obstacles is a fundamental task in robotics that has been modeled in various ways. In Safe Interval Path Planning, location is discretized to a grid, time is continuous, future trajectories of obstacles are assumed known, and planning takes place offline. In this work, we define the Real-time Safe Interval Path Planning problem setting, in which the agent plans online and must issue its next action within a strict time bound. Unlike in classical real-time heuristic search, the cost-to-go in Real-time Safe Interval Path Planning is a function of time rather than a scalar. We present several algorithms for this setting and prove that they learn admissible heuristics. Empirical evaluation shows that the new methods perform better than classical approaches under a variety of conditions.

Tunable Suboptimal Heuristic Search

Stephen Wissow — 2024-06-01

Finding optimal solutions to state-space search problems often takes too long, even when using A* with a heuristic function. Instead, practitioners often use a tunable approach, such as weighted A*, that allows them to adjust a trade-off between search time and solution cost until the search is sufficiently fast for the intended application. In this paper, we study algorithms for this problem setting, which we call `tunable suboptimal search'. We introduce a simple baseline, called Speed*, that uses distance-to-go information to speed up search. Experimental results on standard search benchmarks suggest that 1) bounded-suboptimal searches suffer overhead due to enforcing a suboptimality bound, 2) beam searches can perform well, but fare poorly in domains with dead-ends, and 3) Speed* provides robust overall performance.

A-A*pex: Efficient Anytime Approximate Multi-Objective Search

Han Zhang — 2024-06-01

In the multi-objective search problem, a typical task is to compute the Pareto frontier, i.e., the set of all undominated solutions. However, computing the entire Pareto frontier can be very time-consuming, and in practice, we often have limited deliberation time. Therefore, this paper focuses on solving the multi-objective search problem with anytime algorithms, which compute an initial approximate frontier quickly and then work to find more solutions until eventually finding the entire Pareto frontier. Existing work has investigated such anytime algorithms for problem instances with only two objectives. In this paper, we propose Anytime A*pex (A-A*pex), which works with any number of objectives. In each iteration of A-A*pex, it runs A*pex, a state-of-the-art approximate multi-objective search algorithm, to compute more solutions. From one iteration to the next, A-A*pex can either reuse its previous search effort or restart from scratch. Our experimental results show that an A-A*pex variant that mixes reusing its search effort and restarting from scratch yields the best runtime performance. We also show that A-A*pex often computes solutions that collectively approximate the Pareto frontier much better than the solutions found by state-of-the-art multi-objective search algorithms for short deliberation times.

Bi-Criteria Diverse Plan Selection via Beam Search Approximation

Shanhe Zhong — 2024-06-01

Recent work on diverse planning has focused on a two-step setting where the first step consists of generating a large number of plans, and the second step consists of selecting a subset of plans that maximizes diversity. For the second step, previous work has focused on solving a combinatorial optimization problem for diverse subset selection that can be approximated using greedy search. In this work, we propose a flexible, bi-criteria framework for diverse plan selection. Our framework consists of optimizing both quality and diversity, generalizing previous work and providing flexibility to prioritize one objective over the other. We consider two quality and two diversity measures and show that greedy search guarantees an approximation with a constant ratio for certain configurations based on established results in the literature. To allow users to trade off additional computation for better solutions, we introduce a beam search approximation that generalizes the greedy search, and we provide approximation guarantees on the obtained solutions. Finally, we conduct extensive experiments that show that: (1) our flexible bi-criteria framework allows us to obtain solutions of better quality while still maintaining a high degree of diversity; (2) our beam search approximation obtains significant improvement in performance over greedy search and, for a large number of instances, is able to generate solutions that are equal to or better than those obtained by an exact MIP solver with a significantly higher runtime limit.

Hitting Set Heuristics for Overlapping Landmarks in Satisficing Planning

Clemens Büchner — 2024-06-01

Landmarks are a core component of LAMA, a state-of-the-art satisficing planning system based on heuristic search. It uses landmarks to estimate the goal distance by summing up the costs of their cheapest achievers. This procedure ignores synergies between different landmarks: The cost of an action is counted multiple times if it is the cheapest achiever of several landmarks. Common admissible landmark heuristics tackle this problem by underapproximating the cost of a minimum hitting set of the landmark achievers. We suggest to overapproximate it by computing suboptimal hitting sets instead if admissibility is not a requirement. As our heuristics consider synergies between landmarks, we further propose to relax certain restrictions LAMA imposes on the number of landmarks and synergies between them. Our experimental evaluation shows a reasonable increase in the number of landmarks that leads to better guidance when used with our new heuristics.

Novelty Heuristics, Multi-Queue Search, and Portfolios for Numeric Planning

Dillon Z. Chen — 2024-06-01

Heuristic search is a powerful approach for solving planning problems and numeric planning is no exception. In this paper, we boost the performance of heuristic search for numeric planning with various powerful techniques orthogonal to improving heuristic informedness: numeric novelty heuristics, the Manhattan distance heuristic, and exploring the use of multi-queue search and portfolios for combining heuristics.

Efficient Set Dominance Checks in Multi-Objective Shortest-Path Algorithms via Vectorized Operations

Carlos Hernández Ulloa — 2024-06-01

In the multi-objective shortest-path problem (MOSP) we are interested in finding paths between two vertices of a graph while considering multiple objectives. A key procedure, which dominates the running time of many state-of-the-art (SOTA) algorithms for MOSP is set dominance checks (SDC). In SDC, we are given a set X of N-dimensional tuples and a new N-dimensional tuple p and we need to determine whether there exists a tuple q in X such that q dominates p (i.e., if every element in q is lower or equal than the corresponding element in p). In this work, we offer a simple-yet-effective approach to perform SDC in a parallel manner, an approach that can be seamlessly integrated with most SOTA MOSP algorithms. Specifically, by storing states in memory dimension-wise and not state-wise, we can exploit vectorized operations offered by ``Single Instruction/Multiple Data'' (SIMD) instructions to efficiently perform SDC on ubiquitous consumer CPUs. Integrating our approach for SDC allows to dramatically improve the runtime of existing MOSP algorithms.

Some Orders Are Important: Partially Preserving Orders in Top-Quality Planning

Michael Katz — 2024-06-01

The ability to generate multiple plans is central to using planning in real-life applications. Top-quality planners generate such sets of top-cost plans, allowing flexibility in determining equivalent plans. In terms of the order between actions in a plan, the literature only considers two extremes -- either all orders are important, making each plan unique, or all orders are unimportant, treating two plans differing only in the order of actions as equivalent. To allow flexibility in selecting important orders, we propose specifying a subset of actions the orders between which are important, interpolating between the top-quality and unordered top-quality computational problems. We explore the ways of adapting partial order reduction search pruning techniques to address this new computational problem and present experimental evaluations demonstrating the benefits of exploiting such techniques in this setting.

Optimal Unlabeled Pebble Motion on Trees

Pierre Le Bodic — 2024-06-01

Given a tree, a set of pebbles initially stationed at some nodes of the tree and a set of target nodes, the Unlabeled Pebble Motion on Trees problem (UPMT) asks to find a plan to move the pebbles one-at-a-time from the starting nodes to the target nodes along the edges of the tree while minimizing the number of moves. This paper proposes the first optimal algorithm for UPMT that is asymptotically as fast as possible, as it runs in a time linear in the size of the input (the tree) and the size of the output (the optimal plan).

A Data Efficient Framework for Learning Local Heuristics

Rishi Veerapaneni — 2024-06-01

With the advent of machine learning, there have been several recent attempts to learn effective and generalizable heuristics. Local Heuristic A* (LoHA*) is one recent method that instead of learning the entire heuristic estimate, learns a "local" residual heuristic that estimates the cost to escape a region. LoHA*, like other supervised learning methods, collects a dataset of target values by querying an oracle on many planning problems (in this case, local planning problems). This data collection process can become slow as the size of the local region increases or if the domain requires expensive collision checks. Our main insight is that when an A* search solves a start-goal planning problem it inherently ends up solving multiple local planning problems. We exploit this observation to propose an efficient data collection framework that does <1/10th the amount of work (measured by expansions) to collect the same amount of data in comparison to baselines. This idea also enables us to run LoHA* in an online manner where we can iteratively collect data and improve our model while solving relevant start-goal tasks. We demonstrate the performance of our data collection and online framework on a 4D (x, y, theta, v) navigation domain.

Speeding Up Dominance Checks in Multi-Objective Search: New Techniques and Data Structures

Han Zhang — 2024-06-01

In multi-objective search, given a directed graph where each edge is annotated with multiple cost metrics, a start state, and a goal state. We are interested in computing the Pareto frontier, i.e., the set of all undominated paths from the start state to the goal state. Almost all multi-objective search algorithms use dominance checks to determine if a search node can be pruned. Since dominance checks are performed in the inner loop of the multi-objective search, they are the most time-consuming part of it. In this paper, we propose (1) two novel techniques to reduce duplicate dominance checks and (2) a simple data structure that enables more efficient dominance checks. Our experimental results show that combining our proposed techniques and data structure speeds up LTMOA*, a state-of-the-art multi-objective search algorithm, by up to an order of magnitude on road network instances.

Scaling Lifelong Multi-Agent Path Finding to More Realistic Settings: Research Challenges and Opportunities

He Jiang — 2024-06-01

Multi-Agent Path Finding (MAPF) is the problem of moving multiple agents from starts to goals without collisions. Lifelong MAPF (LMAPF) extends MAPF by continuously assigning new goals to agents. We present our winning approach to the 2023 League of Robot Runners LMAPF competition, which leads us to several interesting research challenges and future directions. In this paper, we outline three main research challenges. The first challenge is to search for high-quality LMAPF solutions within a limited planning time (e.g., 1s per step) for a large number of agents (e.g., 10,000) or extremely high agent density (e.g., 97.7%). We present future directions such as developing more competitive rule-based and anytime MAPF algorithms and parallelizing state-of-the-art MAPF algorithms. The second challenge is to alleviate congestion and the effect of myopic behaviors in LMAPF algorithms. We present future directions, such as developing moving guidance and traffic rules to reduce congestion, incorporating future prediction and real-time search, and determining the optimal agent number. The third challenge is to bridge the gaps between the LMAPF models used in the literature and real-world applications. We present future directions, such as dealing with more realistic kinodynamic models, execution uncertainty, and evolving systems.

Fools Rush in Where Angels Fear to Tread in Multi-Goal CBS

Grigorios Mouratidis — 2024-06-01

Research on multi-agent pathfinding (MAPF) has recently shifted towards problem variants that are closer to actual applications. Such variants often include the assignment of multiple goals to agents. To solve them, researchers have extended the Conflict Based Search (CBS) algorithm to multiple goals. This extension might look straightforward at first sight but it is tricky and this has already led to the development of algorithms that despite claiming to be optimal, return suboptimal solutions for some MAPF instances. In this paper, we provide a detailed analysis of the issue to raise awareness among the search community so that this mistake will not be perpetuated. Furthermore, a first evaluation against an optimal implementation is conducted which shows why this issue might have been difficult to spot. In only one of the randomly generated instances, the suboptimal behavior emerged.

Parallelizing Multi-objective A* Search (Extended Abstract)

Saman Ahmadi — 2024-06-01

The Multi-objective Shortest Path (MOSP) problem aims to find all Pareto-optimal paths between two points in a graph with multiple edge costs. Recent studies on multi-objective search with A* have demonstrated superior performance in solving difficult MOSP instances. This paper proposes a novel parallel multi-objective search framework that can accelerate recent A*-based solutions by several factors.

Finding a Small, Diverse Subset of the Pareto Solution Set in Bi-Objective Search (Extended Abstract)

Pablo Araneda — 2024-06-01

Bi-objective search requires computing a Pareto solution set which contains a set of paths. In real-world applications, Pareto solution sets may contain several tens or even hundreds of solutions. For a human user trying to commit to just one of these paths, navigating through a large solution set may become overwhelming, which motivates the problem of computing small, good-quality subsets of Pareto frontiers. This document presents two main contributions. First, we provide a simple formalization of good-quality subsets of a Pareto solution set. For this, we use measure of richness which has been employed in the study of Population Dynamics. Second, we propose Chebyshev BOA*, a variant of BOA* to compute good-quality subset approximations.

Extreme Value Monte Carlo Tree Search (Extended Abstract)

Masataro Asai — 2024-06-01

Monte-Carlo Tree Search (MCTS) combined with Multi-Armed Bandit (MAB) has had limited success in domain-independent classical planning until recently. Previous work (Wissow and Asai 2023) showed that UCB1, designed for bounded rewards, does not perform well when applied to the cost-to-go estimates of classical planning, which are unbounded in R, then improved the performance by using a Gaussian reward MAB instead. We further sharpen our understanding of ideal bandits for planning tasks by resolving three issues: First, Gaussian MABs under-specify the support of cost-to-go estimates as [−∞, ∞]. Second, Full-Bellman backup that backpropagates max/min of samples lacks theoretical justifications. Third, removing dead-ends lacks justifications in Monte-Carlo backup. We use Extreme Value Theory Type 2 to resolve them at once, propose two bandits (UCB1-Uniform/Power), and apply them to MCTS for classical planning. We formally prove their regret bounds and empirically demonstrate their performance in classical planning.

Finiding All Optimal Solutions in Multi-Agent Path Finding (Extended Abstract)

Shahar Bardugo — 2024-06-01

The Multi-Agent Path Finding problem (MAPF) aims to find conflict-free paths for a group of agents leading each agent to its respective goal. In this paper, we study the requirement of finding all optimal solutions in MAPF. We discuss the representation of all optimal solutions, propose three algorithms for finding them, and compare the algorithms experimentally.

Crafting a Pogo Stick in Minecraft with Heuristic Search (Extended Abstract)

Yarin Benyamin — 2024-06-01

Minecraft is a widely popular video game renowned for its intricate environment. The game's open-ended design allows the creation of unique tasks and challenges for the agents, providing a broad spectrum for researchers to experiment with different AI techniques and applications. Indeed, various Minecraft tasks have been posed as an AI challenge. Most AI research on Minecraft focused on either applying Reinforcement Learning (RL) to solve the problem, learning an action model for planning, or modeling the problem for a domain-independent planner. In this work, we focus on the combinatorial search aspect of solving the Craft Wooden Pogo task within the Polycraft World AI Lab (PAL) Minecraft environment. PAL is an interface to Minecraft that provides an API for AI agents to interact with Minecraft's environment and send commands to the main character. PAL supports symbolic observations of the current state, making it ideal for planning algorithms, which require a symbolic model of the environment for problem-solving. Other Minecraft research frameworks such as MineRL, provide a visual, pixel-based representation of the game.

Taming Discretised PDDL+ through Multiple Discretisations (Extended Abstract)

Matteo Cardellini — 2024-06-01

The PDDL+ formalism allows the use of planning techniques in applications that require the ability to perform hybrid discrete-continuous reasoning. PDDL+ problems are notoriously challenging to tackle, and to reason upon them a well-established approach is discretisation. Existing systems rely on a single discretisation delta or, at most, two: a simulation delta to model the dynamics of the environment, and a planning delta, that is used to specify when decisions can be taken. However, there exist cases where this rigid schema is not ideal, for instance when agents with very different speeds need to cooperate or interact in a shared environment, and a more flexible approach that can accommodate more deltas is necessary. To address the needs of this class of hybrid planning problems, in this paper we introduce a reformulation approach that allows the encapsulation of different levels of discretisation in PDDL+ models, hence allowing any domain-independent planning engine to reap the benefits. Further, we provide the community with a new set of benchmarks that highlights the limits of fixed discretisation.

Traffic Flow Optimisation for Lifelong Multi-Agent Path Finding (Extended Abstract)

Zhe Chen — 2024-06-01

Multi-Agent Path Finding (MAPF) is a fundamental problem in robotics that asks us to compute collision-free paths for a team of agents, all moving across a shared map. Existing scalable approaches struggle as the number of agents grows, as they typically plan free-flow optimal paths, which creates congestion. To tackle this issue, we propose a new approach for MAPF where agents are guided to their destination by following congestion-avoiding paths. Empirically, we report large improvements in overall throughput for lifelong MAPF while coordinating more than ten thousand agents.

Exploring Conflict Generating Decisions: Initial Results (Extended Abstract)

Md Solimul Chowdhury — 2024-06-01

Boolean Satisfiability (SAT) is an NP-complete problem, indicating its inherent computational hardness. However, Conflict Driven Clause Learning (CDCL) SAT solvers efficiently tackle large instances in diverse domains. Swift conflict identification is crucial for effective problem-solving, as conflicts lead to the learning of search space pruning clauses, pinpointing the root causes of conflicts and preventing their recurrence. CDCL decision heuristics prioritize variables that participated in recent conflicts, anticipating rapid conflict generation and expediting additional clause learning. In practice, only a fraction of decisions lead to conflicts, yet some decisions may yield multiple conflicts. In this paper, we delve into a detailed study of conflict generating decisions in CDCL, distinguishing between single conflict (sc) decisions, generating only one conflict, and multi-conflict (mc) decisions, producing two or more conflicts. Our empirical analysis characterizes each decision type based on the quality of the learned clauses they produce. Furthermore, our theoretical analysis reveals a crucial distinction: consecutive clauses learned within the same mc decision form a chain of clauses, absent in learned clauses from sc decisions. This leads to the hypothesis that the reasons for conflicts in mc decisions are more closely related than the reasons for conflicts in sc decisions, empirically confirmed with our introduced notion of reason proximity. Finally, we propose score reduction (sr) as a novel decision strategy, reducing the selection priority of certain variables from learned clauses in mc decisions. With four sets of benchmarks, culminating in over 1200 benchmarks, empirical evaluation of sr implemented on top of the SAT competition 2023 winner solver reveals the merit of this new strategy.

Deployable Yet Effective Traffic Signal Optimisation via Automated Planning (Extended Abstract)

Anas El Kouaiti — 2024-06-01

The use of planning techniques in traffic signal optimisation has proven effective in managing unexpected traffic conditions as well as typical traffic patterns. However, significant challenges concerning the deployability of generated signal plans remain, as planning systems need to consider constraints and features of the actual real-world infrastructure on which they will be implemented. To address this challenge, we introduce a range of PDDL+ models embodying technological requirements as well as insights from domain experts. The proposed models have been extensively tested on historical data using a range of well-known search strategies and heuristics, as well as alternative encodings. Results demonstrate their competitiveness with the state of the art.

Lazy Evaluation of Negative Preconditions in Planning Domains (Extended Abstract)

Santiago Franco — 2024-06-01

AI planning technology faces performance issues with large-scale problems with negative preconditions. In this extended abstract, we show how to leverage the power of the Finite Domain Representation (FDR) used by the popular Fast Downward planner for such domains. FDR improves scalability thanks to its use of multi-valued state variables. However, it scales poorly when dealing with negative preconditions. We propose an alternative hybrid approach that evaluates negative preconditions on the fly during search but only when strictly needed. This is compared to the traditional use of domain-specific PDDL bookmark predicates, increasing memory usage, and automated transformations to Positive Normal Form, further escalating memory consumption.

Minimizing State Exploration While Searching Graphs with Unknown Obstacles (Extended Abstract)

Daniel Koyfman — 2024-06-01

We address the challenge of finding a shortest path in a graph with unknown obstacles where the exploration cost to detect whether a state is free or blocked is very high (e.g., due to sensor activation for obstacle detection). The main objective is to solve the problem while minimizing the number of explorations. To achieve this, we propose MXA∗, a novel heuristic search algorithm based on A∗. The key innovation in MXA∗ lies in modifying the heuristic calculation to avoid obstacles that have already been revealed. Furthermore, this paper makes a noteworthy contribution by introducing the concept of a dynamic heuristic. In contrast to the conventional static heuristic, a dynamic heuristic leverages information that emerges during the search process and adapts its estimations accordingly. By employing a dynamic heuristic, we suggest enhancements to MXA∗ based on real-time information obtained from both the open and closed lists. We demonstrate empirically that MXA∗ finds the shortest path while significantly reducing the number of explored states compared to traditional A∗. The code is available at https: //github.com/bernuly1/MXA-Star.

A New Upper Bound for the Makespan of Cost-Optimal Solutions for Multi-Agent Path Finding (Extended Abstract)

Rodrigo López — 2024-06-01

A well-known approach to solving Multi-Agent Path Finding (MAPF) optimally is compilation to Boolean Satisfiability or Answer Set Programming (ASP). Such compilation-based approaches are superior to other approaches on dense, relatively small instances and may invoke the solver multiple times, each with an encoding of the same instance for a different makespan. Critical to their performance is the runtime of the last solver invocation, whose input is the instance encoded with a theoretical upper bound of the makespan of the optimal solution. In this paper, we propose a new theoretical upper bound for such a last invocation. Unlike the previously known bound, when given a MAPF instance P, our bound requires a solution to P_1, a version of P where one of its agents is removed. We prove that our bound is correct and experimentally significantly tighter than the previously known bound. We propose a recursive parallel approach that allows us to exploit our new bound effectively. Our evaluation of warehouses and random MAPF benchmarks of varied sizes shows that our bound is, on average, 21.2% smaller than the previous bound. This allows for generating grounded ASP formulas around 33.45% smaller and solving 4.9% more instances.

Evaluating Distributional Predictions of Search Time: Put Up or Shut Up Games (Extended Abstract)

Sean Mariasin — 2024-06-01

Metareasoning can be a helpful technique for controlling search in situations where computation time is an important resource, such as real-time planning and search, algorithm portfolios, and concurrent planning and execution. Metareasoning often involves an estimate of the remaining search time of a running algorithm, and several ways to compute such estimates have been presented in the literature. In this paper, we argue that many applications actually require a full estimated probability distribution over the remaining time, rather than just a point estimate of expected search time. We study several methods for estimating such distributions, including some novel adaptations of existing schemes. To properly evaluate the estimates, we introduce `put-up or shut-up games', which probe the distributional estimates without requiring infeasible computation. Our experimental evaluation reveals that estimates that are more accurate in expected value do not necessarily deliver better distributions, yielding worse scores in the game.

A Quality Diversity Approach to Automatically Generate Multi-Agent Path Finding Benchmark Maps (Extended Abstract)

Cheng Qian — 2024-06-01

Multi-Agent Path Finding (MAPF) is a complex problem aiming at searching for paths where teams of agents navigate to their goal locations without collisions. Recent advancements in MAPF have highlighted the necessity for robust benchmarks to evaluate their performance. Previously, the benchmarks used to evaluate MAPF algorithms are predominantly fixed, human-designed maps, which cannot evaluate the behavior of the algorithms comprehensively, leading to potential failures in diverse map scenarios. Meanwhile, quality diversity (QD) algorithm is used to generate maps of high solution quality for MAPF. We employ this technique to automatically generate diverse benchmark maps and explore the detailed behavior of MAPF algorithms in the generated maps. As a preliminary result, we concentrate on EECBS, a popular sub-optimal MAPF algorithm, and observe several findings regarding the runtime and solution quality of EECBS, and difficulty of the generated maps.

Spectral Clustering in Rule-based Algorithms for Multi-agent Path Finding (Extended Abstract)

Irene Saccani — 2024-06-01

We address rule-based algorithms for multi-agent path finding (MAPF). MAPF is a task of finding non-conflicting paths connecting agents' initial and goal positions in a shared environment specified via an undirected graph. Rule-based algorithms use a fixed set of predefined primitive operations to move agents to their goal positions in a complete manner. We propose to apply spectral clustering on the underlying graph to decompose the graph into highly connected components and move agents to their goal cluster first before the rule-based algorithm is applied. The benefit of this approach is twofold: (1) the rule-based algorithms are often more efficient on highly connected clusters and (2) we can potentially run the algorithms in parallel on individual clusters.

On Parallel External-Memory Bidirectional Search (Extended Abstract)

Lior Siag — 2024-06-01

Parallelization and External Memory (PEM) techniques significantly enhance the capabilities of search algorithms for solving large-scale problems. While previous research on PEM has primarily centered on unidirectional algorithms, this work presents a versatile PEM framework that integrates both uni- and bi-directional best-first search algorithms.

CoRe Challenge 2022/2023: Empirical Evaluations for Independent Set Reconfiguration Problems (Extended Abstract)

Takehide Soh — 2024-06-01

In this extended abstract, we describe CoRe Challenge 2022/2023, an international competition series aiming to construct the technical foundation of practical research for Combinatorial Reconfiguration. This competition series targets one of the most well-studied reconfiguration problems, called the independent set reconfiguration problem under the token jumping model, which asks a step-by-step transformation between two given independent sets in a graph. Theoretically, the problem is PSPACE-complete, which implies that there exist instances such that even a shortest transformation requires super-polynomial steps with respect to the input size under the assumption of $NP \neq PSPACE$. The competition series consists of four tracks: three tracks take two independent sets of a graph as input, and ask the existence of a transformation, a shortest transformation, a longest transformation between them; and the last track takes only a number of vertices as input, and asks for an instance of the specified number of vertices that needs a longer shortest transformation steps. We describe the background of the competition series and highlight the results of the solver and graph tracks.

Non-Refined Abstractions in Counterexample Guided Abstraction Refinement for Multi-Agent Path Finding (Extended Abstract)

Pavel Surynek — 2024-06-01

Counterexample guided abstraction refinement (CEGAR) represents a powerful symbolic technique for various tasks such as model checking and reachability analysis. Recently, CEGAR combined with Boolean satisfiability (SAT) has been applied for multi-agent path finding (MAPF), a problem where the task is to navigate agents from their start positions to given individual goal positions so that agents do not collide with each other. The recent CEGAR approach used the initial abstraction of the MAPF problem where collisions between agents were omitted and were eliminated in subsequent abstraction refinements. We propose in this work a novel CEGAR-style solver for MAPF based on SAT in which some abstractions are deliberately left non-refined. This adds the necessity to post-process the answers obtained from the underlying SAT solver as these answers slightly differ from the correct MAPF solutions. Non-refining however yields order-of-magnitude smaller SAT encodings than those of the previous approach and speeds up the overall solving process.

Mixed Integer Programming for Time-Optimal Multi-Robot Coverage Path Planning with Efficient Heuristics (Extended Abstract)

Jingtao Tang — 2024-06-01

We investigate time-optimal Multi-Robot Coverage Path Planning (MCPP) for both unweighted and weighted terrains, which aims to minimize the coverage time, defined as the maximum travel time of all robots. Specifically, we focus on a reduction from MCPP to Min-Max Rooted Tree Cover (MMRTC). For the first time, we propose a Mixed Integer Programming (MIP) model to optimally solve MMRTC, resulting in an MCPP solution with a coverage time that is provably at most four times the optimal. Moreover, we propose two suboptimal yet effective heuristics that reduce the number of variables in the MIP model, thus improving its efficiency for large-scale MCPP instances. We show that both heuristics result in reduced-size MIP models that remain complete (i.e., guaranteed to find a solution if one exists) for all MMRTC instances. We validate the effectiveness of our MIP-based MCPP planner through experiments that compare it with two state-of-the-art MCPP planners on various instances, demonstrating a reduction in the coverage time by an average of 27.65% and 23.24% over them, respectively.

Large-Scale Multi-Robot Coverage Path Planning via Local Search (Extended Abstract)

Jingtao Tang — 2024-06-01

We study graph-based Multi-Robot Coverage Path Planning (MCPP) that aims to compute paths for multiple robots to cover all vertices of a given 2D grid terrain graph G. Existing graph-based MCPP algorithms rely on computing a tree cover on G and then employ the Spanning Tree Coverage (STC) paradigm to generate coverage paths on the decomposed graph D of G. In this paper, we take a different approach by exploring how to systematically search for good coverage paths directly on D. We introduce a new algorithmic framework, called LS-MCPP, which leverages a local search to operate directly on D. We propose ESTC, that extends STC to achieve complete coverage for MCPP on any decomposed graphs, even those resulting from incomplete terrain graphs. Furthermore, we demonstrate how to integrate ESTC with three novel types of neighborhood operators into our framework to effectively guide its search process. Remarkably, LS-MCPP scales efficiently to handle MCPP instances with 32 robots on terrain graphs with 11,892 vertices with just minutes of runtime, thereby showcasing its significant benefits for large-scale real-world coverage tasks.

From Space-Time to Space-Order: Directly Planning a Temporal Planning Graph by Redefining CBS (Extended Abstract)

Yu Wu — 2024-06-01

The majority of multi-agent path finding (MAPF) methods compute collision-free space-time paths which require agents to be at a specific location at a specific discretized timestep. However, executing these space-time paths directly on robotic systems is infeasible due to real-time execution differences (e.g. delays) which can lead to collisions. To combat this, current methods translate the space-time paths into a temporal plan graph (TPG) that only requires that agents observe the order in which they navigate through locations where their paths cross. However, planning space-time paths and then post-processing them into a TPG does not reduce the required agent-to-agent coordination, which is fixed once the space-time paths are computed. To that end, we propose a novel algorithm Space-Order CBS that can directly plan a TPG and explicitly minimize coordination. Our main theoretical insight is our novel perspective on viewing a TPG as a set of space-visitation order paths where agents visit locations in relative orders (e.g. 1st vs 2nd) as opposed to specific timesteps. We redefine unique conflicts and constraints for adapting CBS for space-order planning. We experimentally validate how Space-Order CBS can return TPGs which significantly reduce coordination, thus subsequently reducing the amount of agent-agent communication and leading to more robustness to delays during execution.

Optimal and Bounded Suboptimal Any-Angle Multi-agent Pathfinding (Extended Abstract)

Konstantin Yakovlev — 2024-06-01

Multi-agent pathfinding (MAPF) is the problem of finding a set of conflict-free paths for a set of agents. We explore how to solve MAPF problems when each agent can move between any pair of possible locations as long as traversing the line segment connecting them does not lead to a collision with the obstacles. This is known as any-angle pathfinding. We present the first optimal any-angle multi-agent pathfinding algorithm. Our planner is based on the Continuous Conflict-based Search (CCBS) algorithm and an optimal any-angle variant of the Safe Interval Path Planning (TO-AA-SIPP). The straightforward combination of those, however, scales poorly. To mitigate this, we adapt two techniques from classical MAPF to the any-angle setting, namely Disjoint Splitting and Multi-Constraints. Experimental results on different combinations of these techniques show they enable solving over 30% more problems than the vanilla combination of CCBS and TO-AA-SIPP. In addition, we present a bounded-suboptimal variant of our algorithm, that enables trading runtime for solution cost in a controlled manner.

Multi-agent Motion Planning through Stationary State Search (Extended Abstract)

Jingtian Yan — 2024-06-01

Multi-Agent Motion Planning (MAMP) finds various real-world applications in fields such as traffic management, airport operations, and warehouse automation. This work primarily focuses on its application in large-scale automated warehouses. Recently, Multi-Agent Path-Finding (MAPF) methods have achieved great success in finding collision-free paths for hundreds of agents within automated warehouse settings. However, these methods often use a simplified assumption about the robot dynamics, which limits their practicality and realism. In this paper, we introduce a three-level MAMP framework called PSS which incorporates the kinodynamic constraints of the robots. PSS combines MAPF-based methods with Stationary Safe Interval Path Planner (SSIPP) to generate high-quality kinodynamically-feasible solutions. Our method shows significant improvements in terms of scalability and solution quality compared to existing methods.

Multi-Agent Motion Planning with Bézier Curve Optimization under Kinodynamic Constraints (Extended Abstract)

Jingtian Yan — 2024-06-01

Multi-Agent Motion Planning (MAMP) is a problem that seeks collision-free dynamically-feasible trajectories for multiple moving agents in a known environment while minimizing their travel time. In this paper, we introduce a three-level planner called PSB that combines search-based and optimization-based techniques to address the challenges posed by MAMP. PSB fully considers the kinodynamic capability of the agents and produces solutions with smooth speed profiles. Empirically, we evaluate PSB within the domain of obstacle-rich grid map navigation for mobile robots. PSB shows up to 49.79% improvements in solution cost compared to existing methods while achieving significant improvement in scalability.

A Short Summary of Multi-Agent Combinatorial Path Finding with Heterogeneous Task Duration (Extended Abstract)

Yuanhang Zhang — 2024-06-01

Multi-Agent Combinatorial Path Finding (MCPF) seeks collision-free paths for multiple agents from their initial locations to destinations, visiting a set of intermediate target locations in the middle of the paths, while minimizing the sum of arrival times. While a few approaches have been developed to handle MCPF, most of them simply direct the agent to visit the targets without considering the task duration, i.e., the amount of time needed for an agent to execute the task (such as picking an item) at a target location. MCPF is NP-hard to solve to optimality, and the inclusion of task duration further complicates the problem. To handle task duration, we develop two methods, where the first method post-processes the paths planned by any MCPF planner to include the task duration and has no solution optimality guarantee; and the second method considers task duration during planning and is able to ensure solution optimality. The numerical and simulation results show that our methods can handle up to 20 agents and 50 targets in the presence of task duration, and can execute the paths subject to robot motion disturbance.

Planning and Exection in Multi-Agent Path Finding: Models and Algorithms (Extended Abstract)

Yue Zhang — 2024-06-01

In applications of Multi-Agent Path Finding (MAPF), it is often the sum of planning and execution times that needs to be minimised (i.e., the Goal Achievement Time). Yet current methods seldom optimise for this objective. Optimal algorithms reduce execution time, but may require exponential planning time. Non-optimal algorithms reduce planning time, but at the expense of increased path length. To address these limitations we introduce PIE (Planning and Improving while Executing), a new framework for concurrent planning and execution in MAPF. We first show how PIE for one-shot MAPF improves practical performance compared to sequential planning and execution.We then adapt PIE to Lifelong MAPF, a popular application setting where agents are continuously assigned new goals and where additional decisions are required to ensure feasibility. We examine a variety of different approaches to overcome these challenges and we conduct comparative experiments vs. recently proposed alternatives. Results show that PIE substantially outperforms existing methods for One-shot and Lifelong MAPF.

Multi-Robot Coordination and Layout Design for Automated Warehousing (Extended Abstract)

Yulun Zhang — 2024-06-01

With the rapid progress in Multi-Agent Path Finding (MAPF), researchers have studied how MAPF algorithms can be deployed to coordinate hundreds of robots in large automated warehouses. While most works try to improve the throughput of such warehouses by developing better MAPF algorithms, we focus on improving the throughput by optimizing the warehouse layout. We show that, even with state-of-the-art MAPF algorithms, commonly used human-designed layouts can lead to congestion for warehouses with large numbers of robots and thus have limited scalability. We extend existing automatic scenario generation methods to optimize warehouse layouts. Results show that our optimized warehouse layouts (1) reduce traffic congestion and thus improve throughput, (2) improve the scalability of the automated warehouses by doubling the number of robots in some cases, and (3) are capable of generating layouts with user-specified diversity measures. We include the source code at: https://github.com/lunjohnzhang/warehouse_env_gen_public

Arbitrarily Scalable Environment Generators via Neural Cellular Automata (Extended Abstract)

Yulun Zhang — 2024-06-01

We study the problem of generating arbitrarily large environments to improve the throughput of multi-robot systems. Prior work proposes Quality Diversity (QD) algorithms as an effective method for optimizing the environments of automated warehouses. However, these approaches optimize only relatively small environments, falling short when it comes to replicating real-world warehouse sizes. The challenge arises from the exponential increase in the search space as the environment size increases. Additionally, the previous methods have only been tested with up to 350 robots in simulations, while practical warehouses could host thousands of robots. In this paper, instead of optimizing environments, we propose to optimize Neural Cellular Automata (NCA) environment generators via QD algorithms. We train a collection of NCA generators with QD algorithms in small environments and then generate arbitrarily large environments from the generators at test time. We show that NCA environment generators maintain consistent, regularized patterns regardless of environment size, significantly enhancing the scalability of multi-robot systems in two different domains with up to 2,350 robots. Additionally, we demonstrate that our method scales a single-agent reinforcement learning policy to arbitrarily large environments with similar patterns. We include the source code at https://github.com/lunjohnzhang/warehouse_env_gen_nca_public.