论文部分内容阅读
针对过程系统连续变量优化问题中普遍存在的多峰现象，初步探索了应用模拟退火法求解全局最优解的问题，针对变量只有上下限不等式约束的问题，根据连续变量问题的特性，提出了一种相邻状态的产生函数，并分析了模拟退火过程的起始温度、终止温度以及降温速度对优化计算的影响，给出了这些参数的适宜区域，通过三个例题的计算，将模拟退火法与传统优化方法─基于梯度的方法进行了对比分析，结果表明该法能够有效地解决传统的确定型优化方法所不能奏效的全局最优解问题。
Aiming at the multi-peak phenomena prevailing in continuous variable optimization of process systems, the problem of solving the global optimal solution by using simulated annealing method is preliminarily studied. For the problem that the variables have upper and lower bound inequality constraints, according to the characteristics of the continuous variable problem, The generation function of the adjacent states and the influence of the starting temperature, the ending temperature and the cooling rate of the simulated annealing process on the optimization calculation are analyzed. The suitable regions of these parameters are given. The simulated annealing method Compared with the traditional optimization method based on gradient, the results show that this method can effectively solve the global optimal solution problem that the traditional deterministic optimization methods can not work.