间歇过程PSO-SQP混合优化算法研究

PSO-SQP Hybrid optimization algorithm for batch processes

针对SQP算法在求解具有复杂约束的间歇过程优化时容易陷入局部极值点的问题,本文提出一种PSO-SQP混合优化算法。该算法首先采用外点罚函数法将间歇过程有约束的优化问题转换为无约束的优化问题,利用PSO强大的全局搜索能力对其进行求解,并把搜索结果作为SQP搜索初始点,以此弥补SQP全局搜索弱的缺点,再利用SQP良好的局部收敛性和较强的非线性收敛速度对原优化问题进行精细搜索,弥补了PSO局部搜索弱的缺点,通过不断的迭代最终获得优化问题的全局最优解。该算法充分利用了SQP和PSO的优缺点,增强了其对复杂约束优化问题的求解能力。将本文提出的算法用于连续搅拌化学反应系统温度控制中,仿真结果表明产物浓度能够充分逼近期望值,且反应器的温度轨迹收敛,从而验证了该算法的有效性和实用价值。

Since the SQP（ Sequential Quadratic Programming,SQP） algorithm is easy to fall into local extreme points,in solving complex constrained optimization problems of batch processes,a PSO- SQP（ Particle Swarm Optimization,PSO） hybrid optimization algorithm is proposed. Firstly,the algorithm utilizes the powerful global search capability of PSO to solve the new optimization problems,which is transformed from the original constrained optimization problems by the outer point penalty function method. The solution is calculated as a searching initial point of SQP algorithm to compensate for the problem of SQP weak global search. Secondly,the algorithm utilizes the good local convergence performance and strong nonlinear convergence speed of SQP algorithm,which makes a careful search for the optimization problem to make up shortcomings of the PSO weak local search. Finally the algorithm can get the global optimal solution of optimization problems through constant iterative. The algorithm takes full advantage of the characteristics of PSO and SQP to enhance the algorithm solving abilities for complex constrained optimization problems. In order to verify the efficiency,the proposed algorithm is applied to the temperature control of continuous stirred- tank reactor. Simulation results show that the product concentration can approximate the desire value and the temperature trajectory of continuous stirred- tank reactor is convergent. The efficiency and practicability of the algorithm is verified.