Trust-based secure and optimal route selection in MANET utilizing multiple agent-based reinforcement learning

Abstract

A mobile ad-hoc network (MANET) is a wireless network that has a set of moving nodes merged, with no constant infrastructure where nodes are self-configured. Secure routing is essential for preventing the mobile devices from other vulnerabilities, and thus, efficient characteristics in MANET are exploited to obtain an effective secure routing. The existing techniques have the drawbacks of security and network traffic. In this research, trust-based secure and optimal route selection by multiple agent-based reinforcement learning (MA-based RL) is proposed. Initially, optimum routes are selected through using MA-based RL algorithm via a secure communication. The proposed algorithm reduces reliability, packet delivery ratio (PDR), and breaks service. The performance of the developed algorithm is determined through performance measures of PDR, throughput, delay, and energy consumption with several nodes. The proposed algorithm attains high PDR 94.2%, 93.1%, 92.4% and 90.8% for 50, 100, 150, 200 nodes respectively, which is comparatively effective than the previous methods of two-tier security mechanism (TTSM), trust based adaptive genetic algorithm (TAGA), adaptive trust-based secure and optimal route selection utilizing hybrid fuzzy optimization (ATSORS – HFO), and trust-based topology hiding multipath routing protocol (T-TOHIP).