Abstract: Under the background of big data,to select the suitable Web services from substantial candidate services and to compose the composition service which could accomplish complex value added service demand,an Improved Firework Algorithm (IFWA) was proposed and applied to the discrete service composition optimization problem.According to the contribution of fireworks to the mutated spark,Gauss mutation probability was introduced,which made Gauss mutation have the directivity and maintain the population diversity.In the process of selecting fireworks,the elite selection strategy was adopted to reduce the time expenses of algorithm effectively.The feasibility and validity of the proposed algorithm was verified with a large number of experimental results.

Key words: service composition, firework algorithm, quality of service

摘要： 为了在大数据背景下从大量候选服务集中选择出合适的Web服务,并组合成能够完成复杂增值业务过程需求的组合服务,提出一种改进的烟花算法,首次应用于离散服务组合优化问题。该算法根据烟花对变异后的火花产生的贡献大小,引入高斯变异概率,使高斯变异具有方向性的同时保持了种群多样性;在烟花选择过程中采用精英选择策略,有效降低了算法的时间开销。通过大量实验结果验证了该算法的可行性和有效性。

关键词: 服务组合, 烟花算法, 服务质量