基于改进蚁群算法的服务组合研究

在服务计算过程中，服务组合问题是其中关键的技术之一。在原子候选服务数目巨大的情况下，经典的算法一般都是寻找问题的最优解，存在运算量大，运行时间长的缺点，蚁群算法并不是寻找服务组合问题的最优解，而是得到用户能够认同的可行解。为了能够更有效的为用户提供各种服务，在静态的服务组合建立过程中，以服务发现的候选原子服务集合中的服务质量为权重，将服务组合问题分解成一个有向无环图，在组合代价为最小的原则下，采用改进的蚁群算法为搜索方法，迭代一定的次数或者达到用户设定的服务质量为算法的终止条件，找到能够组合为用户需要的原子候选服务集合，进而快速、准确的得到用户期望的服务。

Service Composition Based on Improved Ant Colony Algorithm

In service computing, service composition problem is one of the key technologies. Large number of candidate services in the atomic case, the classical algorithms are generally looking for the optimal solution, there is large amount of computation, the shortcomings of a long running time, ant colony optimization services portfolio problem is not finding the optimal solution, but the user can identify a feasible solution. In order to more effectively provide various services for users, a static portfolio of services in the building process, to serve a collection of atomic services discovered in the candidate quality of service for weight, the service composition problem into a directed acyclic graph, in combination of the principle of minimum costs, an improved ant colony algorithm for the search method, the number of iterations or a certain quality of service to the user to set the termination conditions for the algorithm to find that combination of candidates for service users need a collection of atoms, then fast and accurate service to the user expectations.