Mobile edge computing rate maximization method based on deep reinforcement learning

Abstract

The invention discloses a mobile edge computing rate maximization method based on deep reinforcement learning. The method comprises the following steps: 1) in a wirelessly-powered edge computing system consisting of a base station and a plurality of pieces of wireless equipment, computing the total rate of all the wireless equipment in a system under given mode selection; 2) searching for optimalmode selection through a reinforcement learning algorithm, i.e. mode selection M<0> and M<1> for all the wireless equipment; and 3) using the mode selection M<0> and M<1> for all the wireless equipment as the system state x<t> for reinforcement learning, executing an action a which is the change to the system state x<t>, setting current reward r(x<t>,a) into a positive value if the total computingrate of the changed system is higher than a previous computing rate, otherwise setting the current reward r(x<t>,a) into a negative value, meanwhile making the system enter a next state x<t+1>, and repeating the iterative process until optimal mode selection M<0> and M<1> is reached. Through adoption of the method, the total computing rate of all the wireless equipment is maximized on the premiseof ensuring user experience.

Description

technical field [0001] The invention belongs to the field of communication, and in particular relates to a communication system for mobile edge computing and a method for maximizing the rate of mobile edge computing based on deep reinforcement learning. Background technique [0002] Recent developments in IoT technology are a critical step towards truly intelligent and autonomous control, especially in many important industrial and commercial systems. In an IoT network, a large number of wireless devices (WDs) capable of communication and computing are deployed. Due to device size constraints and production cost considerations, IoT devices (such as sensors) often carry batteries with limited capacity and energy-saving low Performance processors, therefore, limited device lifetime and low computing power cannot support the growing number of sustainable new applications requiring high-performance computing, such as autonomous driving and augmented reality. The deployment of w...

AI Technical Summary

Problems solved by technology

In an IoT network, a large number of wireless devices (WDs) capable of communication and computing are deployed. Due to device size constraints and production cost considerations, IoT devices (such as sensors) often carry batteries with limited capacity and energy-saving low Performance processors, therefore, limited device lifetime and low computing power cannot support the growing number of sustainable new applications requiring high-performance computing, such as autonomous driving and augmented reality

The deployment of wireless power transfer systems (WPT) can solve the two aforementioned performance problems, but frequent device battery failures not only disrupt the normal operation of individual wireless devices but also significantly degrade the overall network performance, for example, in wireless sensor networks Sensing accuracy

Traditional wireless systems require frequent manual battery replacement, which is expensive and inconvenient. Due to strict battery capacity constraints, in battery-powered wireless systems, minimizing energy consumption and prolonging the operating life cycle of wireless devices is a key design

Images