{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,10,22]],"date-time":"2026-10-22T17:04:26Z","timestamp":1792688666866,"version":"4.5.0"},"reference-count":0,"publisher":"International Consortium of Academic Professionals for Scientific Research","issue":"2","license":[{"start":{"date-parts":[[2026,6,2]],"date-time":"2026-06-02T00:00:00Z","timestamp":1780358400000},"content-version":"unspecified","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0"}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["IJGASR"],"abstract":"<jats:p>The growing adjustment capability of cyber attackers poses hurdles to conventional security systems, as opponent nonstop change their strategies to overcome existing defenses. Conventional approaches, like rule-based firewalls and signature-based intrusion detection systems, based on static assumptions and often unsuccessful by evolving threats. Reinforcement learning (RL) present a dynamic option; however, most prior work considers a single learning agent interacting with a fixed competitor.This paper offers an adversarial reinforcement learning framework in which attacker and defender behave as an autonomous representative within a network shows as a structure of graph. Their communication is behave as a zero-sum Markov game to express the competitive nature of cyber conflicts. The model join the probabilistic intrusion detection, attacker trade-offs between stealth and speed, and resource-constrained protective actions, including concise monitoring and non permanent edge blocking. Through repeated interactions system Both agents utilize independent tabular Q-learning to learn optimal policies. In this paper the experimental results explain that the model reached a high attacker success rate of 97.7% with a close-zero mean reward (0.005), shows fixed and successful learning. The proposed system behave rapidly with an average of one step per episode, and success rates enhance from 96.7% in early stages to 98.7% in later stages, demonstrating consistent performance over time. The results shows the effectiveness of incorporating probabilistic detection and resource-aware procedure in realistic cyber defense scenarios, while providing a actual and interpretable statement for future adversarial cyber security research.<\/jats:p>","DOI":"10.55938\/ijgasr.v5i2.443","type":"journal-article","created":{"date-parts":[[2026,6,2]],"date-time":"2026-06-02T10:04:57Z","timestamp":1780394697000},"page":"188-199","source":"Crossref","is-referenced-by-count":12,"title":["Adversarial Reinforcement Learning for Adaptive Cyber Security Defence"],"prefix":"10.55938","volume":"5","author":[{"given":"Deepika","family":"Dubey","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Geetam","family":"Singh Tomar","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Dinesh","family":"Sharma","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"34968","published-online":{"date-parts":[[2026,6,2]]},"container-title":["International Journal for Global Academic &amp; Scientific Research"],"original-title":[],"link":[{"URL":"https:\/\/journals.icapsr.com\/index.php\/ijgasr\/article\/download\/443\/760","content-type":"application\/pdf","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/journals.icapsr.com\/index.php\/ijgasr\/article\/download\/443\/760","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,6,2]],"date-time":"2026-06-02T10:04:58Z","timestamp":1780394698000},"score":1,"resource":{"primary":{"URL":"https:\/\/journals.icapsr.com\/index.php\/ijgasr\/article\/view\/443"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,6,2]]},"references-count":0,"journal-issue":{"issue":"2","published-online":{"date-parts":[[2026,6,2]]}},"URL":"https:\/\/doi.org\/10.55938\/ijgasr.v5i2.443","relation":{},"ISSN":["2583-3081"],"issn-type":[{"value":"2583-3081","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,6,2]]}}}