CAJ | 학술논문

针对标准粒子群算法收敛速度慢和易陷入局部最优的局限性，提出了一种基于仿生学改进的粒子群算法。即通过在标准粒子群公式中加入负梯度项，使算法更加符合鸟群觅食的实际规律，同时使算法的全局和局部搜索能力得到了平衡。仿真对比结果表明，改进的粒子群算法减小了陷入局部极值的可能性，能够提高最优解的精度和优化效率。
침대표준입자군산법수렴속도만화역함입국부최우적국한성，제출료일충기우방생학개진적입자군산법。즉통과재표준입자군공식중가입부제도항，사산법경가부합조군멱식적실제규률，동시사산법적전국화국부수색능력득도료평형。방진대비결과표명，개진적입자군산법감소료함입국부겁치적가능성，능구제고최우해적정도화우화효솔。
The classic Particle Swarm Optimization has some deficiencies, such as falling in the local optimal region, slow convergence velocity, and so on. Aimed at these disadvantages an improved PSO algorithm is proposed. By employing the information about negative gradient to the standard particle swarm algorithm formula, an improved PSO algorithm can make the equilibrium more closed to the real rules of birds swarm’s foraging. At the same time, the global and local search ability of algorithm is balanced. Simulation results show that, an improved PSO algorithm reduces the chances of getting into the local extremum. At the same time, it can improve the solution accuracy of optimal solution and optimizing efficiency.