| Popis: |
Dynamic service scheduling is a challenging problem to the high-efficient implementation of service based business applications. Recent researches on dynamic service scheduling mostly focus on the service allocation, but they seldom optimize the timetable which defines the starting time of each service instance, and the existing methods proposed are time-consuming and with low calculation precision. In this paper, a novel hybrid algorithm for dynamic service scheduling is proposed. Our algorithm can improve the service allocation by the knowledge from self-learning and social learning and make the timetable more properly by reducing idle time of services. Experimental results show our algorithm can solve the dynamic service scheduling problem more precisely and with less execution time than the existing works. |
