Q-Learning-based edge node resource allocation method and system under power distribution Internet of Things

• Main Authors: QIN SIJUN, TAN CHUANYU, ZHANG ZHIMING, WANG PENG, GUO SHEN, BAI SHUAITAO, ZHANG JICHUAN, ZHANG MINGYU, LIN JIAYING, SUN HAOYANG
• Format: Patent
• Language: Chinese; English
• Published: 16.12.2022
• Subjects: ELECTRIC COMMUNICATION TECHNIQUE; ELECTRICITY; INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTEDFOR SPECIFIC APPLICATION FIELDS; PHYSICS; TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHICCOMMUNICATION

The invention discloses a Q-Learning-based edge node resource allocation method and system under a power distribution Internet of Things. The method comprises the following steps: constructing a state space matrix based on a calculation strategy for executing all calculation tasks of an intelligent node; calculating execution energy consumption and time of all tasks under different calculation strategies by sequentially utilizing a pre-constructed intelligent node calculation task execution model based on the state space matrix, and updating Q values corresponding to the calculation strategies based on the execution energy consumption and time; taking the calculation strategy corresponding to the maximum Q value as an optimal calculation strategy, and distributing each calculation task based on the optimal strategy; wherein the calculation strategy comprises the steps of executing a calculation task at the intelligent node or the edge node; and the intelligent node calculation task execution model is construc