This repository is the result of our research in problems falling under the umbrella of "Timed Multi-agent Patrolling" or "Frequency-based Multi-agent Patrolling", i.e. non-adversarial patrolling problems where cooperative agents aim at optimizing a metric based on their visit times.
This repository contains Java code for a simple simulation environment and various patrolling solutions from existing literature, as well as new solutions (proposed as part of this research). Additionally, it includes Python code for a reinforcement learning environment and new solutions for multi-agent patrolling.
See the list of projects and research articles in the sections below.
- GraphExploration: Implementations related to robot-inspired algorithms for exploration and navigation in graph-like environments, as published by Sampaio & Pereira (2019).
- PatrollingStrategies: Implementations of new patrolling strategies (algorithms) proposed in our research (Sampaio, Sousa & Rocha, 2016, 2018; Sampaio & da Silva, 2019).
- PatrollingStrategiesFromLiterature: Implementations of some patrolling strategies (algorithms) proposed by other researchers.
- SearchLibrary: A library for classical search, including implementations of BFS, DFS, A*, SMA*, etc.
- SimplePatrol: The simulator used by all the strategies implemented in Java (from PatrollingStrategies and PatrollingStrategiesFromLiterature).
- SPjMetalEvo: A project with evolutionary algorithms for multi-agent patrolling (Torreao, Sampaio & Miranda, c. 2015).
- Python-RL-Strategies: Implementation of a Reinforcement Learning environment for multi-agent patrolling in Python, along with some new (Deep) Reinforcement Learning solutions.
- Sampaio, Pablo A.; Sousa, R. S.; Rocha, Alessandro N. New patrolling strategies with short-range perception. In: 2016 XIII Latin American Robotics Symposium and IV Brazilian Robotics Symposium (LARS/SBR). IEEE, 2016. p. 157-162.
- Sampaio, Pablo A.; Sousa, R. S.; Rocha, Alessandro N. Reducing the range of perception in multi-agent patrolling strategies. Journal of Intelligent & Robotic Systems, v. 91, p. 219-231, 2018.
- Sampaio, Pablo A.; da Silva, Kenedy F. S. Decentralized Strategies Based on Node Marks for Multi-Robot Patrolling on Weighted Graphs. In: 2019 Latin American Robotics Symposium (LARS), 2019 Brazilian Symposium on Robotics (SBR) and 2019 Workshop on Robotics in Education (WRE). IEEE, 2019. p. 317-322.
- Sampaio, Pablo A.; Pereira, Jonatan W. Multi-Robot Navigation and Exploration in Graphs with Entrance-Dependent Identification of the Edges. In: 2019 Latin American Robotics Symposium (LARS), 2019 Brazilian Symposium on Robotics (SBR) and 2019 Workshop on Robotics in Education (WRE). IEEE, 2019. p. 323-328.
- Torreao, Vitor de A.; Sampaio, Pablo A.; Miranda, Pericles B. C. An Evolutionary Approach to the Timed Multi-Agent Patrolling Problem. Unpublished, written c. 2015.