In this work, a framework for the integration of planning, scheduling and control using data-driven methodologies is proposed. The framework consists of addressing the integrated problem as a grey-box optimization problem, and using data-driven feasibility analysis and surrogate models to approximate the unknown black-box constraints. We follow a systematic procedure to achieve this integration, consisting of two building blocks: first, we address the integration of scheduling and control followed by the integration of planning and scheduling. To handle dimensionality issues, we introduce the concept of feature selection when building the surrogate models. The methodology is applied to the optimization of an enterprise of air separation plants.

在这项工作中，提出了使用数据驱动的方法来集成计划，调度和控制的框架。该框架包括解决作为灰箱优化问题的集成问题，并使用数据驱动的可行性分析和替代模型来近似未知的黑箱约束。我们遵循一个系统的过程来实现此集成，该过程包括两个构建块：首先，我们解决计划和控制的集成问题，然后解决计划和计划的集成问题。为了处理尺寸问题，我们在构建替代模型时引入了特征选择的概念。该方法应用于空分设备企业的优化。