A 2-Competitive Largest Job on Least Loaded Machine Online Algorithm based on Multi Lists Scheduling Model (2001.07061v1)

Abstract: Online scheduling in identical machines with makespan minimization has been a well studied research problem in the literature. In online scheduling, the scheduler receives a list of jobs one by one and assigns each incoming job on the fly irrevocably to one of the machines before the arrival of the next job. In this paper, we study a variant of the online scheduling problem, where jobs are requested on the fly from k distinct sources. On simultaneous arrival of jobs from multiple sources, it is a non-trivial research challenge to ensure fairness in scheduling with respect to distinct sources, while optimizing the overall makespan of the schedule. We develop a novel Multi Lists Scheduling Model(MLS) and propose a 2-competitive deterministic online algorithm namely Largest Job on Least Loaded Machine(LJLLM) based on the MLS model to address the above mentioned research challenge.