RWCE算法控制参数动态更新促进换热网络结构进化策略

Dynamic updating of controlling parameters for enhancing the structure evolution of heat exchanger network with RWCE algorithm

相较其他进化算法,强制进化随机游走(RWCE)算法能够始终保持较高的种群多样性,从而有效地跳出局部最优。然而,目前对于该算法进化过程中的控制参数如最大步长、最小换热量或换热面积以及接受差解概率的设置仍无严格定义,其取值方法和取值范围都将对结构进化的进度、换热单元生成和消去速度以及最终的换热单元数产生直接影响。根据换热单元数设定逐渐变化的控制参数,进行逐级优化尝试。引入logistic函数作为接受差解概率的取值策略,使最大步长和保留系数均随换热单元数线性变化,实现控制参数的动态更新从而促进换热网络结构进化。通过算例验证,该策略能提高RWCE算法优化换热网络的效率,可获得更理想的网络结构。

Compared with the other evolutionary algorithms,random walk with compulsive evolution(RWCE)algorithm remains high population diversity when optimizing heat exchanger network,which effectively helps jump out of local minimum.But for some parameters such as the maximum step length,minimum heat load or heat exchanging area and also the probability of accepting unsatisfactory setting,there is no strict definition.In addition,the value range of these parameters directly affects the process of structure evolution,the speed of generation and elimination of heat exchanging units,and the final number of the heat exchanging unit.In this paper,these parameters were adjusted gradually according to the number of heat exchanging unit with the aim of obtaining optimal design.The logistic function was adopted as the strategy to control the probability of accepting unsatisfactory results.Meanwhile,the maximum step length was linearly changed with the number of heat exchanging unit.The computational results showed that the proposed method could improve the efficiency of RWCE algorithm and had a good capability to find a better network structure.