From 4bc1863cf344123928cf0e943b373b04a5e6d81f Mon Sep 17 00:00:00 2001 From: Yulun Zhang Date: Tue, 17 Feb 2026 11:53:33 -0500 Subject: [PATCH 1/2] add qd-mapper paper; add page for tmo paper; update year to 2026 --- _data/preprints.yml | 19 +-------------- _data/pubs.yml | 51 ++++++++++++++++++++++++++--------------- _pages/publication.html | 2 +- 3 files changed, 35 insertions(+), 37 deletions(-) diff --git a/_data/preprints.yml b/_data/preprints.yml index e4639cc52303..c47456fa894d 100644 --- a/_data/preprints.yml +++ b/_data/preprints.yml @@ -29,21 +29,4 @@ Code: https://github.com/JingtianYan/SMART/ Demo: https://smart-mapf.github.io/demo/ Video: https://youtu.be/irtFxMjyJXs - abstract: "We present Scalable Multi-Agent Realistic Testbed (SMART), a realistic and efficient software tool for evaluating Multi-Agent Path Finding (MAPF) algorithms. MAPF focuses on planning collision-free paths for a group of agents. While state-of-the-art MAPF algorithms can plan paths for hundreds of robots in seconds, they often rely on simplified robot models, making their real-world performance unclear. Researchers typically lack access to hundreds of physical robots in laboratory settings to evaluate the algorithms. Meanwhile, industrial professionals who lack expertise in MAPF require an easy-to-use simulator to efficiently test and understand the performance of MAPF algorithms in their specific settings. SMART fills this gap with several advantages: (1) SMART uses a physics-engine-based simulator to create realistic simulation environments, accounting for complex real-world factors such as robot kinodynamics and execution uncertainties, (2) SMART uses an execution monitor framework based on the Action Dependency Graph, facilitating seamless integration with various MAPF algorithms and robot models, and (3) SMART scales to thousands of robots. In addition, we use SMART to explore and demonstrate research questions about the execution of MAPF algorithms in real-world scenarios. " - - -- key: Qian2024 - title: A Quality Diversity Method to Automatically Generate Multi-Agent Path Finding Benchmark Maps - site: https://airtclick.github.io/Benchmark.github.io/ # project page - authors: [Cheng Qian, Yulun Zhang, Varun Bhatt, Matthew C. Fontaine, Stefanos Nikolaidis, Jiaoyang Li] - equal_contributions: [Cheng Qian, Yulun Zhang] - venue: arXiv - year: 2024 - thumbnail: /files/yulunzhang/thumbnails/qd-mapf-bench.png # Save to /files/[your-folder]/ - eprint: arXiv:2409.06888 # Example: arXiv:2203.02475 (copy from the "cited as" line on the arXiv page) - tags: [mapf] - links: # You can add additional links not listed below - arXiv: https://arxiv.org/abs/2409.06888 - Code: https://github.com/AirTClick/warehouse- - abstract: We use the Quality Diversity (QD) algorithm with Neural Cellular Automata (NCA) to generate benchmark maps for Multi-Agent Path Finding (MAPF) algorithms. Previously, MAPF algorithms are tested using fixed, human-designed benchmark maps. However, such fixed benchmark maps have several problems. First, these maps may not cover all the potential failure scenarios for the algorithms. Second, when comparing different algorithms, fixed benchmark maps may introduce bias leading to unfair comparisons between algorithms. Third, since researchers test new algorithms on a small set of fixed benchmark maps, the design of the algorithms may overfit to the small set of maps. In this work, we take advantage of the QD algorithm to (1) generate maps with patterns to comprehensively understand the performance of MAPF algorithms, (2) be able to make fair comparisons between two MAPF algorithms, providing further information on the selection between two algorithms and on the design of the algorithms. Empirically, we employ this technique to generate diverse benchmark maps to evaluate and compare the behavior of different types of MAPF algorithms, including search-based, priority-based, rule-based, and learning-based algorithms. Through both single-algorithm experiments and comparisons between algorithms, we identify patterns where each algorithm excels and detect disparities in runtime or success rates between different algorithms. - + abstract: "We present Scalable Multi-Agent Realistic Testbed (SMART), a realistic and efficient software tool for evaluating Multi-Agent Path Finding (MAPF) algorithms. MAPF focuses on planning collision-free paths for a group of agents. While state-of-the-art MAPF algorithms can plan paths for hundreds of robots in seconds, they often rely on simplified robot models, making their real-world performance unclear. Researchers typically lack access to hundreds of physical robots in laboratory settings to evaluate the algorithms. Meanwhile, industrial professionals who lack expertise in MAPF require an easy-to-use simulator to efficiently test and understand the performance of MAPF algorithms in their specific settings. SMART fills this gap with several advantages: (1) SMART uses a physics-engine-based simulator to create realistic simulation environments, accounting for complex real-world factors such as robot kinodynamics and execution uncertainties, (2) SMART uses an execution monitor framework based on the Action Dependency Graph, facilitating seamless integration with various MAPF algorithms and robot models, and (3) SMART scales to thousands of robots. In addition, we use SMART to explore and demonstrate research questions about the execution of MAPF algorithms in real-world scenarios. " \ No newline at end of file diff --git a/_data/pubs.yml b/_data/pubs.yml index f4c826febee8..34704de5cc51 100644 --- a/_data/pubs.yml +++ b/_data/pubs.yml @@ -29,6 +29,38 @@ # url: null # abstract: null +############### 2026 ################## + +- key: Qian2026qdmapper + title: "QD-MAPPER: A Quality Diversity Framework to Automatically Evaluate Multi-Agent Path Finding Algorithms in Diverse Maps" + site: https://airtclick.github.io/qdmapper/ # project page + authors: [Cheng Qian, Yulun Zhang, Varun Bhatt, Matthew C. Fontaine, Stefanos Nikolaidis, Jiaoyang Li] + equal_contributions: [Cheng Qian, Yulun Zhang] + venue: AAMAS + volume: null + number: null + pages: null + year: 2026 + thumbnail: /files/yulunzhang/thumbnails/qd-mapf-bench.png # Save to /files/[your-folder]/ + award: null + doi: null # doi is shown in bibtex and used as the link to the publisher site when the publisher field is not defined + publisher: null # link to the publisher; we don't need this if the publisher can be reached by https://doi.org/[doi] + links: # You can add additional links not listed below + arXiv: https://arxiv.org/abs/2409.06888 + Code: https://github.com/AirTClick/QD-MAPPER + Poster: https://drive.google.com/file/d/10Z7I7JoEXJlQvY56viCuY8oh8c0g2zMV/view?usp=drive_link + Slides: https://1drv.ms/b/c/6847b8d033285874/ES114BsqDDpBh6uzmPjAEFsBIIBw1VY-MlWSgpxlgtZCmg?e=d7rO4z + Talk: null + short_version: + venue: SoCS + year: 2024 + pages: 279-280 + url: https://doi.org/10.1609/socs.v17i1.31580 + tags: [mapf] + abstract: We use the Quality Diversity (QD) algorithm with Neural Cellular Automata (NCA) to generate benchmark maps for Multi-Agent Path Finding (MAPF) algorithms. Previously, MAPF algorithms are tested using fixed, human-designed benchmark maps. However, such fixed benchmark maps have several problems. First, these maps may not cover all the potential failure scenarios for the algorithms. Second, when comparing different algorithms, fixed benchmark maps may introduce bias leading to unfair comparisons between algorithms. Third, since researchers test new algorithms on a small set of fixed benchmark maps, the design of the algorithms may overfit to the small set of maps. In this work, we take advantage of the QD algorithm to (1) generate maps with patterns to comprehensively understand the performance of MAPF algorithms, (2) be able to make fair comparisons between two MAPF algorithms, providing further information on the selection between two algorithms and on the design of the algorithms. Empirically, we employ this technique to generate diverse benchmark maps to evaluate and compare the behavior of different types of MAPF algorithms, including search-based, priority-based, rule-based, and learning-based algorithms. Through both single-algorithm experiments and comparisons between algorithms, we identify patterns where each algorithm excels and detect disparities in runtime or success rates between different algorithms. + + + ############### 2025 ################## - key: ZhangMRS25 title: Destination-to-Chutes Task Mapping Optimization for Multi-Robot Coordination in Robotic Sorting Systems @@ -38,7 +70,7 @@ venue: MRS volume: null number: null - pages: null + pages: 1-7 year: 2025 thumbnail: /files/yulunzhang/thumbnails/task-mapping.png award: null @@ -379,23 +411,6 @@ arXiv: https://arxiv.org/abs/2410.01798 -#- key: QianAAAI25 -# title: A Quality Diversity Approach to Automatically Generate Multi-Agent Path Finding Benchmark Maps -# site: null # project page -# authors: [Cheng Qian, Yulun Zhang, Varun Bhatt, Matthew Christopher Fontaine, Stefanos Nikolaidis, Jiaoyang Li] -# equal_contributions: [Cheng Qian, Yulun Zhang] -# venue: AAAI -# pages: null -# year: 2025 -# doi: null -# links: -# arXiv: https://arxiv.org/abs/2409.06888 -# short_version: -# venue: SoCS -# year: 2024 -# pages: 279-280 -# url: https://doi.org/10.1609/socs.v17i1.31580 - - key: WangAAAI25 title: "LNS2+RL: Combining Multi-agent Reinforcement Learning with Large Neighborhood Search in Multi-agent Path Finding" site: null # project page diff --git a/_pages/publication.html b/_pages/publication.html index f2ff633468cf..84ecee674261 100644 --- a/_pages/publication.html +++ b/_pages/publication.html @@ -4,7 +4,7 @@ permalink: /publications/ --- -{% assign years = (2017..2025) | reverse %} +{% assign years = (2017..2026) | reverse %} {% assign highlighted_names = "" %} {% assign display_links = "true" %} From 7ec23fa0de49e6eefe80e766b9824318b5bb5cf5 Mon Sep 17 00:00:00 2001 From: Yulun Zhang Date: Tue, 17 Feb 2026 12:12:31 -0500 Subject: [PATCH 2/2] remove page number of TMO as all pages of MRS start from 1 --- _data/pubs.yml | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/_data/pubs.yml b/_data/pubs.yml index 34704de5cc51..9841b9790916 100644 --- a/_data/pubs.yml +++ b/_data/pubs.yml @@ -70,7 +70,7 @@ venue: MRS volume: null number: null - pages: 1-7 + pages: null year: 2025 thumbnail: /files/yulunzhang/thumbnails/task-mapping.png award: null