Anton Andreychuk

About

Anton Andreychuk is a leading researcher in artificial intelligence and robotics, specializing in heuristic search, multi-agent pathfinding (MAPF), and motion planning under uncertainty. His work bridges classical planning algorithms with modern learning-based methods. Notably, Andreychuk introduced **TransPath** (2023, 19 citations), a pioneering framework that leverages Transformers to learn heuristics for grid-based pathfinding, significantly outperforming traditional metrics like Manhattan distance. He has made foundational contributions to **safe-interval path planning (SIPP)** and its bounded-suboptimal variants, enabling efficient robot navigation among dynamic obstacles. His research on **multi-agent pickup and delivery** problems (2020, 7 citations) and **kinematic-constrained MAPF** (2020, 5 citations) addresses real-world challenges in automated warehouses and robotics. Andreychuk also developed **POGEMA** (2024, 5 citations), a benchmark platform for cooperative multi-agent pathfinding that integrates reinforcement learning. With over 70 total citations across his top-cited works, his impact is evident in both theoretical advances and practical robotics applications. His work on stochastic environments and prioritized planning techniques (2018, 7+ citations) further solidifies his reputation as a key innovator in autonomous navigation.

Research Focus

Key Achievements

Top Papers

1. 1
   TransPath: Learning Heuristics for Grid-Based Pathfinding via Transformers

   19 citations · 2023
2. 2
   Pathfinding in stochastic environments: learning <i>vs</i> planning

   10 citations · 2022
3. 3
   Planning and Learning in Multi-Agent Path Finding

   9 citations · 2022
4. 4
   On the Application of Safe-Interval Path Planning to a Variant of the Pickup and Delivery Problem

   7 citations · 2020
5. 5
   Two Techniques That Enhance the Performance of Multi-robot Prioritized\n Path Planning

   7 citations · 2018
6. 6
   Two Techniques That Enhance the Performance of Multi-robot Prioritized Path Planning

   5 citations · 2018
7. 7
   Revisiting Bounded-Suboptimal Safe Interval Path Planning

   5 citations · 2020
8. 8
   Safe Interval Path Planning and Flatness-Based Control for Navigation of a Mobile Robot among Static and Dynamic Obstacles

   5 citations · 2022
9. 9
   Multi-agent Path Finding with Kinematic Constraints via Conflict Based Search

   5 citations · 2020
10. 10
    POGEMA: A Benchmark Platform for Cooperative Multi-Agent Pathfinding

    5 citations · 2024

Key Collaborators

Contact & Links