基于连续性的换热网络整型变量优化策略

换热网络优化是典型的混合整数非线性问题,其整型变量(换热网络的结构)的组合情况对于其优化的走向以及局部最优解的质量具有至关重要的作用.分析了换热网络结构与性能之间的连续性特征;利用最速下降法对换热网络整形变量进行初步优化以确定合适的换热器数目,生成一系列连续变化的换热网络结构;以结构序列作为优化变量,借助换热网络性能连续性原理,将多维整型变量转换为一维连续变量,对整型变量进行启发式的连续性搜索.通过具体算例求解,该方法帮助优化过程跳出了局部最优的陷阱,并得到了较好的优化结果.

The strategy of integer variable optimization based on the continuity in heat exchanger network

The heat exchanger network optimization is a typical mixed integer nonlinear problem. The integer variable combination plays a crucial role for the path of optimization and the quality of the local optimal solution in the heat exchanger network. In this paper, the steepest descent method was applied to the optimization of integer variables in the heat exchanger network. Based on the continuity between structure and performance, and taking the structure sequence as the optimization variable, the multidimensional integer variable was converted to one dimensional continuous variable and the integer variable was searched continuously for the further optimization in the heat exchanger network. By solving actual cases, the method proposed in this work helped to jump out of the trap of local optimum and get better results in the optimization process.