间歇生产调度混合建模机制与约束处理方法研究

项目名称： 间歇生产调度混合建模机制与约束处理方法研究

项目编号： No.61203109

项目类型： 青年科学基金项目

立项/批准年度： 2013

项目学科： 自动化学科

项目作者： 董明刚

作者单位： 桂林理工大学

项目金额： 26万元

中文摘要： 间歇过程在工业生产被广泛采用，其优化调度对企业节能降耗、提高效益具有重要的意义。生产数据规模大、含噪声、耦合、不确定及调度中的复杂约束，给基于数据的间歇生产调度建模和优化带来了极大的困难。本项目以间歇生产调度为研究对象，结合自适应网络模糊推理系统（ANFIS），对数据与机理建模的结合机制和复杂约束处理方法进行基础性的科学研究。具体内容包括：研究基于数据的ANFIS关键输入变量的选择方法；研究基于瓶颈分解的ANFIS数据建模与机理建模的结合机制；研究基于Oracle罚函数和多目标的复杂约束处理方法，并设计约束差分进化算法；通过仿真研究验证理论和方法的有效性。项目拟建立间歇生产调度混合建模理论，提出新的复杂约束处理方法，以突破现有方法模型实用性不佳、求解困难的局限，为基于数据的间歇生产优化调度方法走向实用奠定理论与技术基础。

中文关键词： 调度；建模；优化算法；Oracle罚方法；

英文摘要： Batch process is widely adopted in industrial production, and its optimal scheduling is of great significance for energy conservation and economic efficiency. However, it is extremely difficult to model and optimize the batch process scheduling based on production data, which include noise, coupling and uncertainty. This project addresses batch production scheduling. Based on the adaptive network fuzzy inference system (ANFIS), we study the hybrid approach of data-based modeling with mechanism modeling and the constraint handling method for complex constraints in batch process scheduling. Firstly, we will study the key input variables selection method. Secondly, we will study the combination approach of ANFIS modeling and mechanism modeling with the help of bottleneck-based decomposition.Thirdly,we will study an effective constraint handling methods based on Oracle penalty and muli-objective ,and propose a new constrainted differential evolution algorithm for batch process scheduling. Finally, we will verify proposed theories and methods through simulation studies. We will setup the theory of hybrid modeling for batch process scheduling and propose a new constrainted handling method for complex constriants.The purpose is to breakthough the limitions of modeling and optimization, and to lay the theoretical and

英文关键词： Scheduling；modeling；optimization algorithm；Oracle penalty method；