- To evaluate and compare the performance of AODV and DSDV routing protocols in Wireless Sensor Networks (WSNs) and analyze how Q-learning improves routing efficiency, energy usage, and network adaptability using metrics such as throughput, packet delivery ratio, delay, and packet drop.
-
NS2 (Network Simulator-2) for protocol simulation
-
TCL Scripts for execution
-
AWK Files for performance analysis
-
NAM Files for visualization
-
XG Files for graph generation
-
Creation of a static 24-node WSN topology
-
Implementation of AODV and DSDV routing protocols
-
Performance evaluation using NS2 trace files
-
Integration of Q-learning into routing
-
Reward computation and Q-table updates
-
Graph generation for delay, throughput, and jitter
-
End-to-end delay comparison
-
Throughput analysis
-
Jitter evaluation
-
Packet delivery and packet drop ratio
-
Routing overhead performance
-
Comparative evaluation of AODV, DSDV, and their Q-learning variants
-
Q-learning enables dynamic route selection through exploration and exploitation.
-
It reduces routing overhead in AODV and improves energy efficiency in DSDV.
-
In static networks, Q-learning offers limited gains since optimal paths already exist.
-
Better adaptability is observed under changing network conditions.
- The study confirms that Q-learning enhances adaptability, reduces overhead, and supports intelligent routing in WSNs. While performance gains in static networks are limited, Q-learning shows strong potential in dynamic environments. Future work suggests extending this approach to mobile networks and applying advanced models like Deep Q-Networks (DQN) for further optimization.