基于迭代粒子群算法的间歇过程优化

Iterative Particle Swarm Algorithm and Its Application to Batch Process Optimization

针对无状态和终端约束的间歇过程动态优化问题,将迭代思想与粒子群优化算法相结合,提出了迭代粒子群算法。算法首先将控制变量离散化,用标准粒子群算法搜索离散控制变量的最优解,并在随后的迭代过程中不断收缩控制变量的搜索域,使优化性能指标和控制轨线不断趋于最优解。为使优化轨线光滑平稳,算法采用三点线性平滑算子对每次迭代结果进行平滑滤波。算法简洁,可行,高效,特别是在系统梯度信息不可得的情况下更具实用性。对一个间歇过程的仿真结果证明了迭代粒子群算法可以有效地解决不含状态和终端约束的间歇过程动态优化问题。

To solve dynamic optimization problems of batch processes without state independent and end - point constraints, an iterative particle swarm algorithm was developed. The main idea of the algorithm was to execute the standard particle swarm optimization iteratively then the control profile would converge to an optimal one. For the method, the control input Was discretized to a finite number of decision variables and particle swarm optimization was then used to search for the best control vector. The searching space contracted as iterations proceeded hence the performance index and control profile could achieve the best value. The results of each iterated calculation were filtered by a three - point linear smooth operator, which makes the optimal trajectory smooth and steady. The algorithm was simple, feasible and efficient. It is especially practical when the system＇ s gradient information is unavailable. The simulation result of a batch process shows that the iterative particle swarm algorithm can solve the dynamic optimization problems effectively if there is no state independent and end - point constraints.