Proceedings of the 20th Conference on Computer Science and Intelligence Systems (FedCSIS)

DOI: http://dx.doi.org/10.15439/2025F7818

Citation: Proceedings of the 20th Conference on Computer Science and Intelligence Systems (FedCSIS), M. Bolanowski, M. Ganzha, L. Maciaszek, M. Paprzycki, D. Ślęzak (eds). ACSIS, Vol. 43, pages 339–344 (2025)

Full text

Abstract. This study introduces the Forest-Inspired Reinforcement Learning (FIRL) algorithm, a novel approach that harnesses the intricate feedback mechanisms observed in forest ecosystems. A multiagent RL system is proposed, where agents maintain mutualistic relationships, exchanging rewards or insights, fostering a cooperative learning environment. The learning process undergoes stages, similar to ecological succession in forests. The initial stages prioritize exploration, while the mature stages emphasize exploitation and refinement. The algorithm incorporates mechanisms to recover from suboptimal decisions, drawing inspiration from a forest's ability to regenerate post disturbances. A dual agent system, inspired by predator-prey dynamics, ensures a balance between exploration and exploitation in the learning process.

References

1. P. Sequeira and M. Gervasio, “Interestingness elements for explainable reinforcement learning: Understanding agents’ capabilities and limitations,” Artif. Intell., vol. 288, p. 103367, 2019.
2. S. Kelly and M. Heywood, “Emergent solutions to high-dimensional multitask reinforcement learning,” Evolutionary Computation, vol. 26, pp. 347–380, 2018.
3. R. Zhang, F. Torabi, L. Guan, D. Ballard, and P. Stone, “Leveraging human guidance for deep reinforcement learning tasks,” 2019.
4. H. Zhang, J. Wang, Z. Zhou, W. Zhang, Y. Wen, Y. Yu, and W. Li, “Learning to design games: Strategic environments in reinforcement learning,” 2017.
5. V. Saggio, B. Asenbeck, A. Hamann, T. Strömberg, P. Schiansky, V. Dunjko, N. Friis, N. Harris, M. Hochberg, D. Englund, S. Wölk, H. Briegel, and P. Walther, “Quantum speed-ups in reinforcement learning,” 2021.
6. X. Wang, J. Zhang, W. Huang, and Q. Yin, “Planning with exploration: Addressing dynamics bottleneck in model-based reinforcement learning,” ArXiv, vol. abs/2010.12914, 2020.
7. S. Gershman, “Deconstructing the human algorithms for exploration,” Cognition, vol. 173, pp. 34–42, 2018.
8. Q. ming Fu, Q. Liu, H. Luo, and J. Chen, “Single trajectory learning: Exploration vs. exploitation,” 2016.
9. T. C. Blanchard and S. J. Gershman, “Pure correlates of exploration and exploitation in the human brain,” bioRxiv, 2017.
10. C. P. Reyer, N. Brouwers, A. Rammig, B. W. Brook, J. Epila, R. F. Grant, M. Holmgren, F. Langerwisch, et al., “Forest resilience and tipping points at different spatio-temporal scales: approaches and challenges,” Journal of Ecology, vol. 103, no. 1, pp. 5–15, 2015.
11. A. Ali, Y. Hafeez, S. M. Hussainn, and M. U. Nazir, “Bio-inspired communication: A review on solution of complex problems for highly configurable systems,” 2020 3rd Int. Conf. on Computing, Mathematics and Engineering Technologies (iCoMET), 2020.
12. M. Mellal and E. Williams, “A survey on ant colony optimization, particle swarm optimization, and cuckoo algorithms,” 2018.
13. D. Kumar, S. Kumar, R. Bansal, and P. Singla, “A survey to nature inspired soft computing,” Int. J. Inf. Syst. Model. Des., vol. 8, pp. 112–133, 2017.
14. B. Kovács, F. Tinya, and P. Ódor, “Stand structural drivers of microclimate in mature temperate mixed forests,” Agricultural and Forest Meteorology, vol. 234, pp. 11–21, 2017.
15. F. Tinya, B. Kovács, A. Bidló, B. Dima, I. Király, G. Kutszegi, F. Lakatos, Z. Mag, S. Márialigeti, et al., “Environmental drivers of forest biodiversity in temperate mixed forests–a multi-taxon approach,” Science of the Total Environment, vol. 795, p. 148720, 2021.
16. M. Chomel, M. Guittonny-Larchevêque, C. Fernandez, C. Gallet, A. DesRochers, D. Paré, B. G. Jackson, and V. Baldy, “Plant secondary metabolites: a key driver of litter decomposition and soil nutrient cycling,” Journal of Ecology, vol. 104, no. 6, pp. 1527–1541, 2016.