چكيده به لاتين
With the emergence of a new generation of wireless communication networks, the need for high data rates and timely receiving of information, especially in real-time applications, has increased dramatically. The use of millimeter wave frequency (mmWave) and terahertz waves is a promising solution to meet the growing traffic demands in wireless networks. Although these technologies can provide higher data rates to provide different services, their use makes the system face challenges. In other words, communications based on millimeter waves and terahertz waves are sensitive to physical obstacles in signal propagation, and despite providing a very high data transfer rate, they experience rapid changes due to obstruction and multipath propagation. One of the ways to overcome these limitations is the use of intelligent reflective surfaces (RIS), which has attracted a lot of attention in recent research. Based on the unique design of these surfaces, its elements can reflect, refract, absorb or focus incoming waves in any desired direction. With these features, RIS is recognized as a new technology in next-generation communication networks to overcome the challenges in millimeter wave (mmWave) propagation, including path attenuation and blocking. Nowadays, with the emergence of time-sensitive and real-time applications, the freshness of information is very important. In order to measure the freshness of the data, a new criterion called Age of information (AoI) has been introduced, which is defined as the time elapsed since receiving the last successful packet at the destination. AoI is modeled as a random method, and features that monitor it, such as time average AoI, maximum average, etc., are used in a period of time to measure efficiency. Also, due to the dynamics of the operational environment of wireless networks, it is difficult or impossible to obtain complete or statistical information in real operational environments. For this purpose, the use of model-free methods can be a suitable solution for the problems of operational environments. In this thesis, the problem of optimizing the average total AoI in multi-user wireless networks based on mmWave and equipped with RIS is investigated, in which several users measure the physical dynamics of their environment and use the access mechanism. Time division multiplexing (TDMA) transmits the measured values in the form of a data packet to the base station. In the following, due to the random nature of the problem and the lack of availability of complete or statistical information from the system, a model-free algorithm based on deep reinforcement learning called Deep Deterministic Policy Gradient (DDPG) is used to solve the problem so that the timing can be Users and their transmitted power, phase shift in RIS and received beam shaping at the base station were calculated. Finally, the performance of the proposed method is examined in terms of convergence characteristics and changes of various parameters in the network, and the results are compared with other methods and networks without RIS. The obtained results show that the performance of the proposed method compared to other approaches will reduce the average total AoI of users.
