An Improved Chaotic Function Based Hapsogwo-Rl Algorithm for Intelligent Vanet Routing Optimization

Abstract

Vehicular ad hoc Networks (VANETs) work under maximal dynamic conditions where high topology changes, non-constant mobility patterns and unsteady wireless links make efficient routing a big challenge in VANETs. Standard optimization routing techniques and the presently used Hybrid PSO-GWO techniques provide performance intensification but still face problems like slow response, limited exploration, premature convergence, insufficient self-learning ability. Therefore, in order to inscribe these limitations, this paper introduces an advanced Chaotic hybrid adaptive PSO-GWO integrated with Reinforcement Learning named as (CHAPSOGWO-RL). This proposed methodology unites the exploitation strength of particle swarm optimization (PSO), the exploration strength of grey wolf optimization (GWO), where the diversity is introduced via chaotic maps, adaptive decision-making capability, and the power of reinforcement learning (RL). In this paper chaotic function are applied to HAPSOGWO-RL, where chaotic functions like logistic, tent, and gaussian maps that generates randomness that widens the search space and reduces stagnation in dynamic networks. RL is another powerful technique that fine-tunes parameters like inertia weight, convergence rate, routing overhead, and acceleration coefficient. The proposed technique is compared with traditional protocols and swarm-based protocols like the ROANCO algorithm, genetic algorithms and K-means clustering algorithms. Then the results show that the proposed method is better in optimization, like improved delivery ratio, high throughput, minimum overhead, delay, and packet loss. CHAPSOGWO-RL is an intelligent routing engine capable of predicting connectivity, breakage, stable forwarders etc. This algorithm is implemented in NS2.35 with SUMO based vehicular traces. Simulation results show that the proposed methodology improves the throughput, increases the delivery ratio, lowers delay, and lowers the packet loss. Overall, proposed algorithm represents a highly adaptive and self-evolving routing framework designed for VANET applications, chaos theory reinforcement learning, and the strengths of swarm intelligence.