Q-balance: An Approach for Balancing Data Imputation Tasks on Edge resources of a Smart Grid.

Smart grids integrate intelligence, automation, and communication into the electrical grid infrastructure, primarily through the use of smart meters. These meters play a crucial role in collecting and transmitting data, either to the cloud, which may cause delays, or to the edge, where meters are closer to the data source. In this paper, we propose Q-Balance, a neural network-based solution for optimizing computational resources at the edge, thus minimizing service processing time. Q-Balance utilizes the Multi-Layer Perceptron (MLP) technique to estimate response times for requests processed by computational resources. Evaluation results demonstrate that Q-Balance can significantly reduce the average response time, achieving up to a 65% reduction compared to the Min-Load approach at the edge and up to 79% in the cloud.