Ing. Kristýna Janovská

  • Profil
  • Publikace

Publikace

Multi-Agent Path Finding in Continuous Environment.

Rok
2024
Publikováno
Proceedings of the 2024 IEEE 36th International Conference on Tools with Artificial Intelligence (ICTAI). Los Alamitos: IEEE Computer Society, 2024. p. 708-713. ISSN 1082-3409. ISBN 979-8-3315-2724-2.
Typ
Stať ve sborníku
Anotace
We address a variant of multi-agent path finding in continuous environment (SC-MAPF), where agents move along sets of smooth curves. Collisions between agents are resolved via avoidance in the space domain. In this work a new Continuous Environment Conflict-Based Search (CE-CBS) algorithm is proposed. CE-CBS combines conflict-based search (CBS) for the high-level search framework with RRT* for low-level path planning. The CE-CBS algorithm is tested under various settings on various SC-MAPF instances. Experimental results show that CE-CBS is competitive w.r.t. other algorithms that consider the continuous aspect in MAPF such as MAPF with continuous time.

Spectral Clustering in Rule-based Algorithms for Multi-agent Path Finding (Extended Abstract).

Autoři
Janovská, K.; Saccani, I.; Surynek, P.
Rok
2024
Publikováno
Proceedings of the Seventeenth International Symposium on Combinatorial Search (SoCS 2024). Menlo Park: AAAI Press, 2024. p. 281-282. ISSN 2832-9163. ISBN 978-1-57735-891-6.
Typ
Stať ve sborníku
Anotace
We address rule-based algorithms for multi-agent path finding (MAPF). MAPF is a task of finding non-conflicting paths connecting agents' initial and goal positions in a shared environment specified via an undirected graph. Rule-based algorithms use a fixed set of predefined primitive operations to move agents to their goal positions in a complete manner. We propose to apply spectral clustering on the underlying graph to decompose the graph into highly connected components and move agents to their goal cluster first before the rule-based algorithm is applied. The benefit of this approach is twofold: (1) the rule-based algorithms are often more efficient on highly connected clusters and (2) we can potentially run the algorithms in parallel on individual clusters.

Agent-Based Modeling in Hierarchical Control of Swarms During Evacuation

Rok
2023
Publikováno
SN Computer Science. 2023, 4 ISSN 2662-995X.
Typ
Článek
Anotace
We address the problem of evacuation from the perspective of agent-based modeling (ABM) in this paper. The evacuation problem is modeled as a navigation of multiple agents that spatially interact with each other in a known environment. The environment is divided into a danger and a safe zone while the task of agents is to move from the danger zone to the safe one in a collision-free manner. Unlike previous approaches that model the environment as a discrete graph with agents placed in its vertices, at most one agent per vertex, our approach adopts various continuous aspects such as a grid-based embedding of the environment into 2D space and continuous line of sight of agents. In addition to this, we adopt hierarchical structure of multi-agent system in which so called leading agents are more informed having full knowledge of other agents and are capable of performing multi-agent pathfinding (MAPF) via centralized algorithms like conflict-based search (CBS) while so called follower agents with limited knowledge about other agents are modeled using simple local rules. Our experimental evaluation indicates that suggested hierarchical modeling approach can serve as a tool for studying the progress and the efficiency of evacuation processes in different environments.

Spectral Clustering in Rule-Based Algorithms for Multi-Agent Path Finding

Autoři
Saccani, I.; Janovská, K.; Surynek, P.
Rok
2023
Publikováno
Proceedings of the 20th International Conference on Informatics in Control, Automation and Robotics. Madeira: SciTePress, 2023. p. 258-265. ISSN 2184-2809. ISBN 978-989-758-670-5.
Typ
Stať ve sborníku
Anotace
We focus on rule-based algorithms for multi-agent path finding (MAPF) in this paper. MAPF is a task of finding non-conflicting paths connecting agents’ specified initial and goal positions in a shared environment specified via an undirected graph. Rule-based algorithms use a fixed set of predefined primitives to move agents to their goal positions in a complete manner. We propose to apply spectral clustering on the underlying graph to decompose the graph into highly connected component and move agents to their goal cluster first before the rule-based algorithm is applied. The benefit of this approach is twofold: (1) the algorithms are often more efficient on highly connected clusters and (2) we can potentially run the algorithms in parallel on individual clusters.

Combining Conflict-based Search and Agent-based Modeling for Evacuation Problems (Extended Abstract)

Rok
2022
Publikováno
Proceedings of the Fifteenth International Symposium on Combinatorial Search. Palo Alto, California: Association for the Advancement of Artificial Intelligence (AAAI), 2022. p. 294-296. vol. 15. ISBN 978-1-57735-873-2.
Typ
Stať ve sborníku
Anotace
We address the problem of evacuation from the heuristic search perspective combined with agent-based modeling (ABM). The evacuation problem is modeled as a navigation of multiple agents in a known environment. The environment is divided into a danger and a safe zone while the task of agents is to move from the danger zone to the safe zone in a collision-free manner. Unlike previous approaches that model the environment as a discrete graph with agents placed in its vertices, at most one agent per vertex, our approach adopts various continuous aspects such as a grid-based embedding of the environment into 2D space and continuous line of sight of agents. In addition to this, we adopt hierarchical structure of our multi-agent system in which so called leading agents are more informed and are capable of performing multi-agent pathfinding (MAPF) via centralized algorithms like conflict-based search (CBS) while so called following agents with limited knowledge about other agents are modeled using simple local rules. Our experimental evaluation indicates that suggested hierarchical modeling approach can serve as a tool for studying the progress and the efficiency of evacuation processes in different environments.

Hierarchical Control of Swarms during Evacuation

Rok
2021
Publikováno
Proceedings of the 13th International Joint Conference on Knowledge Discovery, Knowledge Engineering and Knowledge Management. Setùbal: SciTePress, 2021. p. 61-73. vol. 2. ISSN 2184-3228. ISBN 978-989-758-533-3.
Typ
Stať ve sborníku
Anotace
The problem of evacuation is addressed from the perspective of agent-based modeling (ABM) in this paper. We study evacuation as a problem of navigation of multiple agents in a known environment. The environment is divided into a danger and a safe zone while the task of agents is to move from the danger zone to the safe one. Unlike previous approaches that model the environment as a discrete graph with agents placed in its vertices our approach adopts various continuous aspects such as grid-based embedding of the environment into 2D space continuous line of sight of an agent. In addition to this, we adopt hierarchical structure of multi-agent system in which so called leading agents are more informed and are capable of performing multiagent pathfinding (MAPF) via centralized algorithms like conflict-based search (CBS) while so called follower agents are modeled using simple local rules. Our experimental evaluation indicates that suggested modeling approach can serve as a tool for studying the progress and the efficiency of the evacuation process.