Multivariate data management method and system based on integrated wind and light energy storage

The invention relates to the technical field of data processing, and provides a multivariate data management method and system based on integrated wind and light energy storage, and the method comprises the steps: introducing a first energy storage equipment source data set example without priori learning knowledge, and carrying out the training and optimization through a deep decision tree network, the objective of the invention is to generate an accurate and efficient energy supply optimization strategy. Meanwhile, the second energy supply optimization decision network is debugged in a multi-task learning mode, so that the network training cost is further reduced, and the optimization performance is improved. 本发明涉及数据处理技术领域，提供了一种基于集成风光储能的多元数据管理方法及系统，该方法通过引入不具有先验学习知识的第一储能设备源数据集示例，并利用深度决策树网络进行训练和优化，旨在生成准确高效的能源供应优化策略。同时，本发明还通过多任务学习的方式，对第二能源供应优化决策网络进行调试，以进一步降低网络训练代价并提高优化性能。