Intelligent electric vehicle charging and discharging decision-making method based on deep reinforcement learning
A smart electric vehicle and reinforcement learning technology, applied in neural learning methods, market forecasting, instruments, etc., can solve problems such as delay, difficult modeling, complex modeling, etc., to solve the problem of overestimation, overcome noise and inaccurate The effect of stabilizing and enhancing generalization ability
Patent Information
- Authority / Receiving Office
- CN · China
- Current Assignee / Owner
- Publication Date
- 2021-11-09
Smart Images

Figure 1 
Figure 2 
Figure 3
Abstract
Description
technical field
[0001] The invention relates to the technical field of charging and discharging of electric vehicles, in particular to a decision-making method for charging and discharging of intelligent electric vehicles based on deep reinforcement learning. Background technique
[0002] With the improvement of residents' living standards and the intelligence level of electric vehicles, people will consider more factors when charging and discharging electric vehicles, such as the cumulative charging and discharging costs and user satisfaction. For many families with electric vehicles, if the electric vehicle is charged at the maximum power when it is parked, although the user satisfaction is relatively high, it will bring a large charging cost at the same time; Charging, although it can reduce charging costs, will reduce user satisfaction if the electric vehicle is not fully charged when it leaves.
[0003] Different from other controllable loads and energy storage devices...
Examples
Embodiment Construction
[0037] The specific implementation manners of the present invention will be further described in detail below in conjunction with the accompanying drawings and embodiments. The following examples are used to illustrate the present invention, but are not intended to limit the scope of the present invention.
[0038] A smart electric vehicle charging and discharging decision-making method based on deep reinforcement learning, such as figure 1 Shown, specifically include the following steps:
[0039] Step 1: Collect electricity price data for the past 24 hours;
[0040] Step 2: Use the single-step prediction LSTM network to iteratively predict the electricity price in the next 24 hours, such as figure 2 shown;
[0041] Step 2.1: Expand the LSTM network into a 23-layer neural network, and use the same weight parameters for each layer;
[0042] Step 2.2: Let the input d of the first layer t-22 =p t-22 -p t-23 , where p t-22 and p t-23 Represent the electricity prices at t...