Kei18 / lacam

LaCAM: Search-Based Algorithm for Quick Multi-Agent Pathfinding (AAAI-23)

Home Page:https://kei18.github.io/lacam/

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

lacam

MIT License CI

The code repository of the paper "LaCAM: Search-Based Algorithm for Quick Multi-Agent Pathfinding" (AAAI-23).

Building

All you need is CMake (≥v3.16). The code is written in C++(17).

First, clone this repo with submodules.

git clone --recursive https://github.com/Kei18/lacam.git
cd lacam

Then, build the project.

cmake -B build && make -C build

Docker

You can also use the docker environment (based on Ubuntu18.04) instead of the native one.

# ~10 min, mostly for CMake build
docker compose up -d
docker compose exec dev bash
> cmake -B build && make -C build

Usage

build/main -i assets/random-32-32-10-random-1.scen -m assets/random-32-32-10.map -N 50 -v 1

The result will be saved in build/result.txt.

Output File

This is an example output of random-32-32-10-random-1.scen. (x, y) denotes location. (0, 0) is the left-top point. (x, 0) is the location at x-th column and 1st row.

agents=50
map_file=random-32-32-10.map
solver=planner
solved=1
soc=1316
soc_lb=1113
makespan=55
makespan_lb=53
sum_of_loss=1191
sum_of_loss_lb=1113
comp_time=1
seed=0
starts=(11,6),(29,9),[...]
goals=(7,18),(1,16),[...]
solution=
0:(11,6),(29,9),[...]
1:(10,6),(29,10),[...]
[...]

You can find details of all parameters with:

build/main --help

Visualizer

@Kei18/mapf-visualizer is available.

Experiments

The experimental script is written in Julia ≥1.7. Setup may require around 10 minutes.

sh scripts/setup.sh

Edit the config file as you like. Examples are in scripts/config . The evaluation starts by following commands.

julia --project=scripts/ --threads=auto
> include("scripts/eval.jl"); main("scripts/config/mapf-bench.yaml")

Notes

  • The grid maps and scenarios in assets/ are from MAPF benchmarks.
  • The empirical data of the manuscript was obtained with [exp/AAAI2023].
  • LaCAM with different design choices: see [pilot/greedy] and [pilot/dbs]
  • tests/ is not comprehensive. It was used in early developments.
  • Auto formatting (clang-format) when committing:
git config core.hooksPath .githooks && chmod a+x .githooks/pre-commit

Licence

This software is released under the MIT License, see LICENSE.txt.

Author

Keisuke Okumura is a Ph.D. student at Tokyo Institute of Technology, interested in controlling multiple moving agents.

About

LaCAM: Search-Based Algorithm for Quick Multi-Agent Pathfinding (AAAI-23)

https://kei18.github.io/lacam/

License:MIT License


Languages

Language:C++ 94.1%Language:CMake 3.6%Language:Dockerfile 1.7%Language:Shell 0.6%