隔代强制进化遗传算法在换热网络优化中应用

The Application of Atavistic Forced-evolution Genetic Algorithms in the Optimization of Heat Exchange Networks

在对换热网络分级超结构及其数学模型分析的基础上,对网络综合优化问题进行了研究。针对普通遗传算法及其它优化算法无法保证网络优化质量和效率的缺点,对遗传算法进行了改进,提出了换热网络隔代强制进化遗传算法。该方法将换热网络结构信息转化为种群中染色体信息,利用种群的进化实现网络结构的优化,在进化过程中使用隔代强制策略,使种群向更优方向稳步进化,保证各代优化结果的有效性,降低最优群体的生成代数,并利用最优个体保存技术记录优化过程中最佳换热网络结构。采用此方法对具体换热网络实例进行了优化综合,结果表明隔代强制进化遗传算法能在网络优化过程中避免早熟收敛而陷入局部最小点的现象,使搜索质量和效率得到有效提高。用隔代强制进化遗传算法对有分流和无分流换热网络进行优化综合,均能获得综合性能良好的网络结构。

Based on graded superstructures, the authors have studied the synthetic optimization of heat exchange networks. Through an improvement of the genetic algorithm, presented is an atavistic forced-evolution genetic algorithm for a heat exchange network. By using this method, a specific heat-exchange network is subject to a synthetic optimization. The results of the optimization show that the forced-evolution genetic algorithm can avoid a localized minimum point phenomenon caused by a premature convergence, making it possible to effectively enhance the searching quality and efficiency. The use of a forced-evolution genetic algorithm to synthesize the heat exchange network can result in a network structure possessing a good comprehensive performance.