化工过程综合问题MINLP算法中整型变量的连续化

化工过程系统综合问题需要同时考虑设备结构参数和工艺操作参数，一般用整型变量表示设备的取舍，用连续变量表示操作参数，这就构成一个流程的超结构，在数学形式表现为一个混合整型非线性规划（MINLP）问题。混合整型非线性规划问题的求解成为化工过程综合优化的关键。今根据超结构中整型变量的特征，提出整型变量连续化处理的思路，将MINLP问题简化为NLP问题，然后采用罚函数法求解。最后将该算法运用于加氢脱烷基化（HDA）过程综合的实例研究，结果表明该算法克服了传统方法在处理整型变量时出现的麻烦，为有效快速地进行化工过程综合优化问题提供了一种新的途径。

The Integer Variables Continuation Technique Used in MINLP Problem for Chemical Process Synthesis

The chemical process synthesis problem involves selecting the optimal flowsheet structures and operating parameters to gain the maximum economic profit. In general, the flowsheet structures adopted or not are expreseed by integer variables and the operation parameters are expressed as continuous variables, and these form a superstructure model. In order to simultaneously optimize the operation parameters and flowsheet structure of the chemical process, the superstructure model formulated is generally a MINLP (mixed integer nonlinear programming) problem. How to solve the above model is the key to optimize the chemical process. A new method which is called the integer variables continuation technique was proposed, which is based on the character of the binary variables (the integer variables and the coutinuous variables), they can be treated as mere continuous variables by adding some constrains, and then the difficulties brought by the binary variables in the process of optimization can be avoided. Thereby the optimum structure and the optimum operation parameters are determined in a easier way. The HDA process was taken as a case for studying the capabilities of the proposed method. The result proves that the proposed method can treat complex MINLP problems properly.