Optimizing Charging Strategies for Electric Vehicles in Residential Areas Based on Leaped Halton Sequence Initialized Flood Algorithm

The rapid growth of electric vehicles (EVs) has led to a surge in charging demand, resulting in power grid overload and insufficient charging facilities, thus, it is urgent to optimize models to solve the problem of supply-demand imbalance. At present, relevant research has focused on proposing valuable strategies to solve the problem of charging overload during peak hours. However, these models are based on static prediction and fail to effectively consider the various factors of real-time changes in EV charging, resulting in inflexible system scheduling, congestion during charging peak hours, and poor performance when charging demand fluctuates. And for many charging optimization problems, they are equivalent to non-smooth and non-convex optimization problems, and traditional subgradient descent (SubGD) and coordinate descent (CD) methods are generally prone to getting stuck in local optimal solutions. In order to overcome the above shortcomings, in this paper, a derivative-free optimization method called leaped Halton sequence initialize flood algorithm (LHSFLA) has been proposed, which can break out of the trap of local optimal solutions and find global optimal solution, so as to balance power load fluctuations and user charging costs, and effectively solve the problems of power grid overload and insufficient charging facilities.