Abstract: A new Particle Swarm Optimization (PPSO) algorithm is proposed for solving constrained optimization problems. A feasibility-based rule is used for updating the individual and global best solutions. Adaptive perceptive ability is assigned to particles in PSO for balancing their global and local searching and avoiding prematurity. The velocity of particle around the boundary is revised by the results of perceiving to enhance the searching around the boundary. Simulation results show that the proposed approach has fast convergence and good optimization ability, and is suitable for solving constrained optimization problems.

Key words: constrained optimization problem, particle swarm optimization, adaptive perceptive ability, constrained boundary

摘要： 提出一种求解约束优化问题的改进粒子群优化算法。它利用可行性判断规则处理约束条件，更新个体最优解和全局最优解。通过为粒子赋予自适应感知能力，算法能较好地平衡全局和局部搜索，且有能力跳出局部极值，防止早熟。边界附近粒子的感知结果被用来修正其飞行速度以加强算法对约束边界的搜索。实验结果表明，新算法收敛速度快,寻优能力强，能很好地求解约束优化问题。

关键词: 约束优化问题, 粒子群优化, 自适应感知能力, 约束边界