engleski

A method of service optimization by self-trained path selection

This paper presents an application of reinforcement Q-learning in solving the problem of automatic routing. A potential redundancy has been detected in a telecommunication network comprising multiple nodes, specialized for routing the jobs between the clients and service providers. The main idea presented here is a permanent trans-fer adaptation for the jobs that have to be sent from a client to service providers, throughout network nodes dedicated for routing. The search for optimal path, i. e. self-adaptation is based on continuous monitoring of available channel capacity between the client and network. The actions that follow monitoring are focused on the selection of optimal throughput via the node that enables maximal exploitation the channel capacity. In some situations information flow between the client and service provider can be significantly reduced as a consequence of traffic load and the choice of throughput. From the client view point accumulation and exploration of knowledge about throughput properties in the network can optimally utilize redundant capacities. Given outcomes are the results of computer simulations.

reinforcement learning; self-adaptivity; routing

nije evidentirano