Abstract: To realize process mining and to overcome the disadvantages that Standard Particle Swarm Optimization (SPSO) was easily getting into pre-maturity and local optimum, mutation operation based particle swarm process mining was proposed. In evolutionary process of SPSO, all particles followed the optimal particle to search, which led population diversity to appear premature convergence after decreased rapidly. Under genetic algorithm inspiration, the algorithm was flung off restraint of falling in local extreme point by carrying out mutation operation for particle in population. Thus algorithm's ability to extricate escape from the local optimum was improved. Experimental results showed that the proposed method got good effect in accuracy and speed of solution.

Key words: particle swarm optimization algorithm, process mining, premature convergence, mutation

摘要： 为实现过程挖掘，克服标准粒子群算法易陷入局部极值的缺点，提出基于变异操作的粒子群过程挖掘方法。在标准粒子群算法进化中，所有粒子追随最优粒子在解空间搜索，导致种群多样性迅速下降，出现早熟收敛。受遗传算法启发，通过对进化中的粒子增加变异操作，使算法摆脱易于陷入局部极值点的束缚，增强算法跳出局部最优的能力。仿真结果表明，基于变异粒子群算法的过程挖掘在求解的精度和速度方面都得到了好的效果。

关键词: 粒子群优化算法, 过程挖掘, 早熟收敛