Optimal scheduling method of low-carbon energy supply chain based on machine learning and game theory

Aiming at the quality and stability of a low-carbon new energy supply chain, this paper puts forward an optimal scheduling method for a low-carbon energy supply chain based on machine learning and game theory. By establishing an index and preprocessing keywords, the similarity between the index and the data in the platform database is calculated, and the data-sharing result of the low-carbon energy supply chain is obtained. This paper analyzes the game benefits of each subject in the optimization of the low-carbon energy supply chain, obtains the game matrix, and realizes the optimal scheduling of the low-carbon energy supply chain. The experimental results show that the scheduling convergence error of the low-carbon energy supply chain is low, the accuracy is high, and the load supply equipment can run stably.


Introduction
With the rapid development of the global economy, the pressure on energy demand is increasing.At the same time, the problems of the environment and climate change have also triggered the urgent demand for low-carbon energy [1].As a sustainable energy form, low-carbon energy has become an important direction of global energy transformation because of its environmental friendliness and little impact on climate [2].In the traditional energy supply chain, suppliers, manufacturers, and distributors obtain the maximum benefits through game theory and have achieved remarkable results in practice [3][4].However, in the low-carbon energy supply chain, due to its particularity, the traditional game theory model and optimization method are facing great challenges.As an algorithm technology that can automatically learn and improve, machine learning can analyze and predict a large amount of data in a low-carbon energy supply chain, thus providing data-based decision support [5].Game theory can work out the optimal decision-making scheme by analyzing the strategy and interest relationship among supply chain participants.Machine learning can provide accurate predictions and recommendations for decision-making by analyzing a large number of supply chain data.Therefore, this paper proposes an optimal scheduling method for a low-carbon energy supply chain based on machine learning and game theory.By studying the application of machine learning and game theory in the low-carbon energy supply chain, we hope to provide decision support and theoretical guidance for the development of the low-carbon energy industry and promote the sustainable development of energy transformation.

 
12 , , , where a  stands for keyword index; A stands for document; n stands for term.We match the generated index with the data in the platform database [8] and calculate the similarity between the index and the data in the platform database.The calculation formula is: where W D stands for keyword index [9]; s N stands for the document; v V stands for the term.We match the generated index with the data in the platform database and calculate the similarity between the index and the data in the platform database.The calculation formula is: where Κ represents the expectation of data access times in the low-carbon energy supply chain; ij V stands for the membership degree of low-carbon energy supply chain data [10].Only when the ij  value is greater than zero, it means that the documents in the database are consistent with the index. output value are used as retrieval results and sent to the platform display interface.

Game matrix calculation
To simplify the model, this paper analyzes the game benefits of each subject in the optimization of the low-carbon energy supply chain and obtains the game matrix, as shown in Figure 1.

Subject normal request
Accept the request As can be seen from Figure 1, the game matrix is usually composed of normal requests and abnormal requests of the subject, which is used to describe the game relationship between different participants in resource allocation and decision-making.We suppose that there are two main participants in the low-carbon energy supply chain: energy suppliers and energy consumers.Energy suppliers can choose to provide different types of energy products, while energy consumers can choose to buy energy from suppliers.Their goal is to get the best results by maximizing their profits or utility.By analyzing the game matrix, best strategy combination can be determined so that the participants can achieve the best result.

Realization of optimal scheduling of low-carbon energy supply chain
We collect the data on energy use and carbon emission in the supply chain and make a detailed analysis, including the data of all links from the beginning of energy production to the consumption process, such as energy source, transshipment mode, conversion, and delivery efficiency.We input raw data and set parameters such as data size, conversion rate, mutation rate, and maximum repetition times.Then we calculate the corresponding relationship of all data, that is, the value of each data objective function, and set targets and key performance indicators of the low-carbon energy supply chain.Finally, we establish a mathematical model by using machine learning and optimization algorithms to evaluate the decision-making scheme.The previous generation population selection, hybridization, and mutation operations were carried out to create, replace, and preserve individuals with high adaptability in the new generation population.Therefore, we evaluate the carbon footprint of the energy system, determine the main emission sources and potential emission reduction opportunities, and output the optimal solution of the optimal scheduling and end of the low-carbon energy supply chain.

Experimental analysis
To test the feasibility and reliability of the optimal scheduling method of a low-carbon energy supply chain based on machine learning and game theory, a low-carbon energy supply chain is selected as the experimental object, and the data scheduling method of this low-carbon energy supply chain is used.The methods in [3] and [4] are selected as the comparison.The experimental configuration includes a desktop computer with Windows 2010 operating system, 16 GB hard disk, and i812.25 CPU, and the parameters during the experiment are set, as shown in Table 1.  1, the relevant parameters needed for this experiment are established.On this basis, the generated high-dimensional feature vector matrix is sorted out through the role of the machine learning model and saved in TXT file format.
Before the experiment, this paper chooses a test system, including an energy station, conventional unit, thermoelectric unit, gas supply station, wind farm, and other scenarios.Electric heating prices and natural gas prices in different regions have different changes.In this paper, the low-carbon energy supply chain is optimally scheduled, which has gone through 100 iterations, and the scheduling optimization iteration is recorded every 20 iterations.The result of the scheduling convergence error is shown in Figure 2. As can be seen from Figure 2, the convergence error of system energy scheduling in the initial stage exceeds 0.06, which indicates that there is a big deviation between the scheduling scheme and the actual scheduling environment.With the increase of iterations, after 20 iterations of scheduling optimization, the scheduling convergence error is reduced to 0.038, which means that the scheduling scheme gradually approaches the requirements of the actual scheduling environment.When the number of iterations reaches 100, the scheduling convergence error is about 0.01, which can meet the needs of the actual scheduling environment.This shows that the scheduling scheme of the low-carbon energy supply chain can be continuously improved and optimized through iterative optimization.With the increase in iteration times, the scheduling scheme gradually converges, which is closer to the actual scheduling requirements.Through continuous optimization and adjustment, we can find a better scheduling strategy and improve the efficiency and accuracy of the energy supply chain.In practical application, the system can be scheduled according to the experimental results, to achieve better energy utilization, energy saving, and emission reduction.
An accuracy test is an evaluation method to measure the accuracy of the classification model in the prediction task.In the classification task, the number of samples whose prediction results are consistent with the actual labels accounts for the proportion of the total number of samples.The accuracy is the number of true positives and true negatives divided by the total number of samples.For the energy supply chain scheduling strategy, the higher the accuracy is, the better the strategy effect will be.After the test system completes the scheduling convergence, the optimal scheduling accuracy of the low-carbon energy supply chain is obtained under the condition of generating power of 600 MW, and the result is shown in Figure 3.As can be seen from Figure 3, with the extension of time, the optimal scheduling accuracy of the low-carbon energy supply chain gradually declines, but it gradually maintains a stable accuracy of 96.1% during the decline.This is because establishing indexes and preprocessing keywords helps realize data sharing and integration of a low-carbon energy supply chain.By standardizing the data and establishing an index to speed up the retrieval, the data of different links can be effectively integrated and shared.In the process of scheduling, the data of each link can be comprehensively utilized to improve the accuracy of scheduling decisions.It shows that the scheduling optimization method designed in this paper can meet the conditions of optimal scheduling income of a low-carbon energy supply chain.This method is applied to the scheduling optimization of the low-carbon energy supply chain, and the scheduling performance of this method is analyzed through the optimal control results of system load.The experimental results are shown in Figure 4. From the analysis of Figure 4, it can be seen that to achieve the goal of the lowest operating cost of the comprehensive energy system, the output of wind turbines and power grid purchase are taken as the main ways to realize the load supply of the comprehensive energy system, and the energy storage equipment is enabled to store electric energy to meet the load demand during peak hours.The electric conversion mode is more economical, and the electric boiler can be dispatched to provide the required load.With the increasing demand for various loads, the micro gas turbine is called to operate as the main load supply equipment of the low-carbon energy system in normal time, and the electric boiler and heat storage tank are started as supplementary equipment to meet the load demand in the system.

Conclusion
This study proposes an optimal scheduling method for a low-carbon energy supply chain based on machine learning and game theory.By studying the application of machine learning and game theory in this field, the following conclusions are drawn: (1) After 20 iterations of scheduling optimization, the scheduling convergence error is reduced.When the number of iterations reaches 100, the scheduling convergence error is about 0.01, which can meet the actual scheduling environment.
(2) Using the scheduling optimization method designed in this paper, the optimal scheduling accuracy of the low-carbon energy supply chain can reach 96.1%, which can meet the optimal scheduling income conditions of the low-carbon energy supply chain.
(3) With the increasing demand for various loads, the electricity price calls for the micro gas turbine to run in normal time as the main load supply equipment of the low-carbon energy system, and starts the electric boiler and heat storage tank as supplementary equipment to meet the load demand in the system.
However, there are still some limitations in this study, and the optimal scheduling method of a lowcarbon energy supply chain based on machine learning and game theory needs a lot of data support.Therefore, in practical applications, data collection and analysis may face some challenges.The future research direction can focus on the following aspects.First of all, we can further study the selection and optimization of machine learning algorithms and game theory models to improve the accuracy and stability of the model.Secondly, we can consider introducing other modern technologies, such as the Internet of Things and artificial intelligence, to further improve the management level and intelligence of the low-carbon energy supply chain.

Figure 2 .
Figure 2. Convergence error results from low-carbon energy supply chain scheduling.

Figure 3 .
Figure 3. Optimal scheduling results of low-carbon energy supply chain.

Figure 4 .
Figure 4. Results of load scheduling optimization.