Optimization of Concurrency Coefficients Based on Gaussian Process Regression for Smart Grid Network Bandwidth Estimation

With the rapid development of power communication networks, the bandwidth allocation of a certain site has become the focus of research. The current method of bandwidth calculation is to utilize the formula provided by the national grid file, including four parameters including concurrency coefficient. However, the method of determining the concurrency coefficient is mainly determined by empirical formulas, which does not perform very well. This paper proposes a method for optimizing the concurrent coefficients based on Gaussian process regression. The experiment shows that the error of the result obtained by the optimized concurrency coefficient is smaller than the original one.