Explanation Generation for Multi-Modal Multi-Agent Path Finding with Optimal Resource Utilization using Answer Set Programming

Abstract

The multi-agent path finding (MAPF) problem is a combinatorial search problem that aims at finding paths for multiple agents (e.g., robots) in an environment (e.g., an autonomous warehouse) such that no two agents collide with each other, and subject to some constraints on the lengths of paths. We consider a general version of MAPF, called mMAPF, that involves multi-modal transportation modes (e.g., due to velocity constraints) and consumption of different types of resources (e.g., batteries). The real-world applications of mMAPF require flexibility (e.g., solving variations of mMAPF) as well as explainability. Our earlier studies on mMAPF have focused on the former challenge of flexibility. In this study, we focus on the latter challenge of explainability, and introduce a method for generating explanations for queries regarding the feasibility and optimality of solutions, the nonexistence of solutions, and the observations about solutions. Our method is based on answer set programming.

Flexible and Explainable Solutions for Multi-Agent Path Finding Problems

“…With these motivating real life scenarios, we introduced a method to generate explanations for a variety of queries about mMAPF solutions, using the expressive formalism and efficient solvers of ASP [1]. Our explainable framework is implemented using Python and the ASP solver Clingo.…”

“…Consider the mMAPF instance (Scenario 1 of [1]) shown in Figure 7(a) in a small warehouse, where Robot 1 is initially located at Cell 11 and aims to reach Cell 5, and Robot 2 is initially located at Cell 8 and aims to reach Cell 2. Cell 7 is a waypoint for both robots, and the upper bound on makespan of a plan is 4.…”

“…Once an alternative solution, Plan 2 described in Figure 7(c), is found, our algorithm presents the following explanation: "Actually, Robot 2 does not have to wait at Cell 8 from time step 0 to 2. Here is an alternative optimal plan: Plan 2..." Consider the instance (Scenario 6 of [1]) in Figure 8(b). Robots 1-2 are initially located at 1 and 3; their goals are at 3 and 1, respectively.…”

“…The algorithm further generates the following explanation: "There is no solution because Robot 2 collides with the obstacle at Cell 2 at time step 1; this suggests removing this obstacle." Details of the algorithm and the ASP formulations, together with more examples and experimental evaluations are presented in our paper, Explanation Generation for Multi-Modal Multi-Agent Path Finding with Optimal Resource Utilization using Answer Set Programming [1].…”

DPLL(MAPF): an Integration of Multi-Agent Path Finding and SAT Solving Technologies

“…Compiling MAPF to other formalism for which an off-the-shelf solver is available is a popular solving approach. Optimal solvers for MAPF based on the compilation to constraint satisfaction problem (CSP) [22], answer set programming (ASP) [5], integer programming (IP) [14], and Boolean satisfiability (SAT) [31] currently exist.…”