A hybrid neurodynamics algorithm for solving distributed nonconvex nonsmooth load sharing problems in power systems

In this paper, a hybrid neurodynamics optimizatio (HNO) algorithm is proposed, which is a distributed convex optimization algorithm combined with stochastic neurodynamics. The state solution of the HNO algorithm can avoid falling into local optimal solution by adding neural white noise. The results of numerical experiments show that HNO algorithm has better global optimization ability. In addition, we build a distributed load share (DLS) model. Different from the current model, which only focuses on the quadratic economic cost function, we not only consider the valve point effect and the environmental cost function of the fuel generator, but also consider the constraints of the transmission line which connects the bus. A distributed non-convex nonsmooth optimization problem is constructed. To solve this problem, we first smoothed each local objective function, then dealt with nonconvex constraints utilizing penalty function, and transformed the original problem into a distributed nonconvex optimization problem. Finally, the HNO algorithm is used to solve the problem. The optimization values of each variable in the power grid show that the power grid can operate in an optimal state under the condition of power limitation.