Q-learning based adaptive subgraph generation algorithm for graph routing in ISA 100.11a

Graph routing is a routing protocol opted to run in ISA 100.11a based networks. Attention towards proposing a time efficient subgraph is received by the researchers. While there can be a lot of network performance level factors considered to generate subgraphs for efficient and timely delivery of data. ISA 100.11a specification doesn´t specifies a graph generation algorithm to generate subgraphs for a network. Hence, numerous efforts can be found where the research community has taken initiative in proposing subgraphs considering factors affecting the performance of the network. Subgraphs which are optimal at a particular instance of time may not prove to be effective after a certain time period. Therefore, there is a need to update the subgraphs in the network based on the history of the network performance parameters accumulated over a certain time period which will not necessarily regenerate the subgraphs in the network based on the history will penalize the subgraphs which degrade the performance of the applications using the subgraph. Q-learning mechanism is a lightweight mechanism which works on the idea of rewarding an action considered good for environment. We propose an idea of generating subgraphs after a time period based on Q-learning mechanism.