Cuckoo Search Algorithm for Job Scheduling in Cloud Systems

摘要

Cloud systems are known as a large scale computer system that responds a large number of incoming tasks over the Internet. The system contains running high performance computers on network that merges all computational machines into one system. In this paper, we introduce a Cuckoo search (CS) algorithm to schedule computational jobs efficiently considering system resource constraint. The scheduling problem is considered as an NP-hard problem and the challenge is to distribute tasks over the system for minimizing completion time and execution cost. In order to evaluate the effectiveness and performance of the proposed algorithm, a set of problem instances has been generated and tested with CS and Genetic algorithm (GA). Experimental results show that CS outperforms GA in terms of calculation time and fitness value.