基于自适应粒子群算法的重油热解模型参数估计

通过构造一个合适的目标函数，将化工模型参数估计问题转化为一个多维数值优化问题，然后提出一种参数自适应调整和维变异的改进粒子群优化算法来求解该问题。该算法首先利用佳点集方法初始化种群以保证粒子的多样性。惯性权重和学习因子随进化过程自适应调整，从而协调算法的全局和局部搜索能力。为了避免算法陷入局部最优，对收敛度最小的维进行变异。几个标准测试问题的实验结果表明该算法具有较强的全局寻优能力。最后将改进粒子群算法应用到重油热解模型参数估计中，并与基本遗传算法（SGA）和粒子群优化算法（SPS0）进行比较。研究结果表明：本文得到的平均相对误差为5．62％，比SGA和SPSO分别低1．08％和0．50％。

Parameter Estimation for Heavy Oil Thermal Cracking Model Based on Particle Swarm Optimization Algorithm

Through establishing an appropriate objective function, the parameter estimation problem for chemical engineering model was formulated as a multi-dimensional numerical optimization problem, which can be solved by modifying particle swarm optimiza-tion （MPSO） algorithm with adaptive parameters and dimensional mutation. In this approach, good point set method is used to con-struct the initialization population which strengthens the diversity of global searching. In the evolution process, in order to balance the global and local search abilities of the algorithm, adaptive inertia weights and acceleration coefficients strategies are introduced respectively. Furthermore, dimensional mutation operator is utilized to avoid the premature convergence. Several benchmark func-tions are tested; the experimental results show that the MPSO algorithm is an effective way for global optimization problems. The MPSO is successfully applied to the parameter estimation of the thermal cracking model for heavy oil. The results indicate that the mean relative error of the proposed method is only 5.62 %, which is less than those of standard genetic algorithm and standard parti- cle swarm optimization by 1.08% and 0.51%, respectively.