Trust-driven both-matched algorithm for grid task multi-objective scheduling

Job scheduling algorithm is the key technique in grid management. In order to meet the requirements of high quality QoS in job scheduling of OGSA grid, trust mechanism is added to make the job scheduling system work more effectively in this open, heterogeneous and dynamic grid environment. Utilizing the concept of trust, a trust-driven both-matched grid job scheduling algorithm(TD-BM) is presented, which takes four objective namely system security, system reliability, load balance and system throughput into consideration. Afterwards TD-BM is compared with the traditional TD-Min-min under the same hypothetical condition and the results illustrate that, under the grid environment of high heterogeneous jobs and machines, the scheduling effect of TD-BM is obviously better than TD-Min-min. While keeping almost the same trust value, TD-BM not only shortens the scheduling makespan, but also improves system load balance.