基于知识改进的文化算法及其在化工动态优化中的应用

针对化工以及生化过程的动态优化问题,提出了一种基于改进知识引导的文化算法。该算法首先对控制搜索域与时间域分别进行了等分和离散化,利用软约束思想编码控制序列,采用种群产生-控制域进化-种群寻优迭代过程实现对控制序列的逐步寻优;其次在种群空间采用遗传算法,在信度空间采用差分算法,并将进化过程中的已有种群信息设计为3种知识,通过分析知识、提取知识、管理知识来指导进化过程。由于引入了文化进化理念和机制,大大提高了动态优化问题的搜索效率。通过3种典型化工动态优化问题的仿真实例,表明该算法具有较好的寻优效率以及更好的优化结果,验证了该算法在解决具有非线性动态约束问题的有效性。

Dynamic optimization in chemical processes using improved knowledge-based cultural algorithm

An improved knowledge-based cultural algorithm（IKBCA） was proposed to solve dynamic optimization problem of chemical and biochemical processes. The first step of IKBCA was to discrete time region and control region, and to encode chromosome by taking advantage of soft constraint strategy. The second step was to use the process of iteration: “initialize population”-“evolve control region”-“optimize population” to achieve the optimal control profile. IKBCA inserted genetic algorithm (GA) into the population space and differential evolution algorithm (DE) into the brief space. By analyzing, extracting, and managing three kinds of improved knowledge developed from the individual information computed, the evolutionary process could be well guided. When applied to dynamic optimization problem of three typical chemical processes with distinguishing control features, IKBCA demonstrated a competitive optimal searching ability, and its feasibility and effectiveness were verified for the dynamic nonlinear constrained optimization problem.