Heterogeneous Network Selection Algorithm Based on Deep Q Learning

In order to adapt to the dynamic changes of the network environment, it is necessary to select the most suitable network for each session to serve the heterogeneous network and achieve network load balancing at the same time. Based on the heterogeneous network composed of PDT and B-TrunC, and based on deep Q learning algorithm, the network selection Markov decision process (NSMDP) is adopted. Based on Markov decision-making process, we establish a network selection problem and propose an algorithm for wireless access network selection in heterogeneous network environment. The algorithm considers not only the load of the network, but also the business attributes of the initiating session, the mobility of the terminal and the location of the terminal in the network. The simulation results show that the algorithm reduces the system blocking rate and achieves the autonomy of network selection.