Abstract: Service-oriented cloud computing environment provides novel ideas for knowledge innovation in manufacturing field. The dynamic combination of Knowledge as a Service（KaaS） is one of the key technologies in the process of knowledge innovation. The virtual and dynamic nature of the service resources in cloud computing puts forward new challenges for Quality of Service（QoS） of composite KaaS. In this paper, an improved Harmony Search algorithm（SLHS） is presented for solving the optimization problem about QoS of composite KaaS in manufacturing field. Using the method of Skyline to initialize harmony memory, it can improve the operation efficiency of algorithm. Using the method of TOPSIS to select the manufacturing KaaS, it can ensure the effectiveness of the solution. Basic Harmony Search（BHS） algorithm is introduced to compare with SLHS in the simulation experiment. The experimental results show that the SLHS algorithm is better than BHS algorithm in solution quality and algorithm performance.

Key words: cloud manufacturing, Knowledge as a Service（KaaS）, harmony search algorithms, optimization composite KaaS

摘要： 面向服务的云计算环境为制造领域的知识创新提供了新的思路。知识即服务的动态组合是知识创新过程中的关键技术之一。云计算服务资源的虚拟性和动态性为组合的知识即服务的服务质量提出了新的挑战。针对制造领域知识即服务组合的服务质量优化问题，提出一种改进的和声搜索算法（SLHS），SLHS算法利用Skyline方法对和声记忆库进行初始化以提高算法的运行效率，并采用理想点法选择制造知识即服务以确保解的有效性。仿真实验中引入了基本和声搜索算法作比较。实验结果表明SLHS算法在解的质量方面和算法性能方面均明显优于基本和声搜索算法。

关键词: 云制造, 知识即服务（KaaS）, 和声搜索算法, 服务组合优化