论文部分内容阅读
针对过程系统优化问题中存在多解、而目前通用高性能实时优化算法均为局部算法的情况,提出了基于禁忌搜索的全局记忆增强型优化方法(global mnemonic enhancement optimization,GMEO)。该方法证明了在一定条件下局部最优解是过程系统参数的连续可微函数,并采用禁忌搜索方法以背景计算的方式获得优化计算经验点上的多解信息。基于这些多解信息,GMEO可为优化计算提供全体局部最优解的逼近值。以这些逼近值为初值,可使局部优化算法达到全局优化的效果,并能有效提高过程系统运行的稳健性。基于原油混合问题的数值仿真试验验证了GMEO方法的有效性。在不过多损失实时性的前提下,优化计算的目标函数值得以显著改善。
Aiming at the existence of multiple solutions in the process system optimization problem and the current general high performance real-time optimization algorithms all being local algorithms, a global mnemonic enhancement optimization (GMEO) method based on tabu search is proposed. This method proves that under certain conditions, the local optimal solution is a continuous differentiable function of the process system parameters, and the tabu search method is used to obtain the multi-solution information of the optimization calculation experience point by the way of background calculation. Based on these multi-solution information, GMEO can provide an approximation of the global optimal solution for the optimization calculation. Using these approximations as initial values, the local optimization algorithm can achieve the effect of global optimization and can effectively improve the robustness of the process system operation. The numerical simulation based on the crude oil mixing problem verifies the effectiveness of the GMEO method. Without losing too much real-time performance, the objective function of optimization calculation can be significantly improved.