论文部分内容阅读
针对粒子群优化算法在处理约束问题时产生的不可行解,引用基于多级罚函数的约束处理方法。为了改进罚函数粒子群算法易早熟、后期收敛慢、易陷入局部最优解的缺点,提出了动态改变惩罚系数的改进粒子群算法。应用于几个经典的测试函数,都在较少的迭代次数内得到了高精度的优化解,验证了算法的有效性。以某一机械零部件的可靠性优化为例,建立了基于改进粒子群算法的可靠性优化设计模型。结果表明:该方法能快速有效地解决可靠性优化设计问题,计算结果明显优于常规的多级罚函数法。
Aiming at the infeasible solution of particle swarm optimization algorithm when dealing with the constraint problem, the constraint processing method based on multi-level penalty function is used. In order to improve the penalty function PSO is easy to premature, post-convergence slow, easy to fall into the shortcomings of the local optimal solution, an improved particle swarm algorithm is proposed to dynamically change the penalty coefficient. Applying to several classical test functions, high-precision optimization solutions are obtained with fewer iterations, which verifies the effectiveness of the algorithm. Taking the reliability optimization of a certain mechanical component as an example, a reliability optimization design model based on improved particle swarm optimization is established. The results show that this method can solve the problem of reliability optimization design quickly and effectively, and the calculation result is obviously better than the conventional multi-level penalty function method.