Distributed Network Simulation
Using Reinforcement Learning

This is a running distributed simulation of using reinforment learning algorithms to perform packet routing in computer network.
It enables comparison between different algorithms in different input pattern and network loads.
please choose some simulation run time parameters:

Network Topology:

6X6 Irregular Grid
Small 3-node demo network

Input Pattern:

Random source and destination
All packets sent to one destination
All packets sent from first node to last (in 6X6 GRID it means from upper left to lower right)

Routing Algorithm:

Bellman-Ford shortest-paths algorithm
QRouting algorithm
DUAL-QRouting algorithm

Lookup Table Implementation:

Hashtable (only local information.more overhead)
Array (some knowledge on network size.more efficient)

Total number of packets sent:

1000 3000 5000 10000 50000

Start counting average time:

From simulation start
After half of the packets were delivered (after the learning phase)

Debug on Nodes:

Yes,indeed No,thank you

Debug on Routing:

Yes,indeed No,thank you

Debug on Results:

Yes,indeed No,thank you

REMARK: simulation running may take some time. Enjoy!