Suboptimal Variants of the Conflict-Based Search Algorithm for the Multi-Agent Pathfinding Problem

Abstract

The task in the multi-agent path finding problem (MAPF) is to find paths for multiple agents, each with a different start and goal position, such that agents do not collide. A successful optimal MAPF solver is the conflict-based search (CBS) algorithm. CBS is a two level algorithm where special conditions ensure it returns the optimal solution. Solving MAPF optimally is proven to be NP-hard, hence CBS and all other optimal solvers do not scale up. We propose several ways to relax the optimality conditions of CBS trading solution quality for runtime as well as bounded-suboptimal variants, where the returned solution is guaranteed to be within a constant factor from optimal solution cost. Experimental results show the benefits of our new approach; a massive reduction in running time is presented while sacrificing a minor loss in solution quality. Our new algorithms outperform other existing algorithms in most of the cases.

Unbounded Sub-Optimal Conflict-Based Search in Complex Domains

“…A comparison of the runtime of our two strategies versus previous state of the art, greedy CBS (GCBS) (Barer et al 2014) is shown in Figure 2. Statistics in the plots are the average runtime over 50 problem instances on 16-connected grids with increasing numbers of agents.…”

“…Conflict-Based Search (CBS) (Sharon et al 2015) is a state-of-the-art algorithm for MAPF used in the context of both unit cost and more complex domains. Sub-optimal variants of CBS have focused on techniques such as focal search and alternate search prioritization schemes (Barer et al 2014;Cohen et al 2018). This work instead focuses on relaxing requirements for optimal constraints in order to achieve further performance gains.…”

Lower and Upper Bounds for Multi-Agent Multi-Item Pickup and Delivery: When a Decoupled Approach is Good Enough (Extended Abstract)

“…To find the collision-free paths of the VRP solution, we used two different MAPF solvers. The first one is based on the Priority-Based Search (PBS) (Ma et al 2019), and the second one on the ECBS (Barer et al 2014). PBS is a suboptimal solver which is known to be very fast.…”

“…We show that when used with a bounded suboptimal MAPF solver, i.e., ECBS (Barer et al 2014), the decoupled approach is able to find near-optimal solutions, which means that using more advanced approaches is not required. However, on instances specifically designed to be hard to solve, ECBS cannot find a solution, and PBS generates solutions with a significantly higher cost than the VRP solver.…”

Optimal and Bounded-Suboptimal Multi-Agent Motion Planning

“…While CBS is not directly applicable to the MAMP problem, we successfully extend its core ideas to the richer MAMP domain. To do so, (1) we adapt (Enhanced) CBS (Barer et al 2014) to efficiently reason with continuous time and arbitrary wait durations; (2) we introduce an efficient implementation of reservation tables using interval maps (Cormen et al 2009); and (3) we develop a novel bounded-suboptimal extension of Safe Interval Path Planning (SIPP) (Phillips and Likhachev 2011), called Soft Conflict Interval Path Planning (SCIPP). On the theoretical side, we justify the completeness, optimality and bounded-suboptimality of our MAMP solvers on state lattices.…”