TITLE:
Minimizing Time in Scheduling of Independent Tasks Using Distance-Based Pareto Genetic Algorithm Based on MapReduce Model
AUTHORS:
Devarajan Rajeswari, Veerabadran Jawahar Senthilkumar
KEYWORDS:
Distributed Systems, Multi-Objective, MapReduce, Optimization, DPGA, NSGA-II
JOURNAL NAME:
Circuits and Systems,
Vol.7 No.6,
May
12,
2016
ABSTRACT: Distributed Systems (DS)
have a collection of heterogeneous computing resources to process user tasks.
Task scheduling in DS has become prime research case, not only due of finding
an optimal schedule, but also because of the time taken to find the optimal
schedule. The users of Ds services are more attentive about time to complete
their task. Several algorithms are implemented to find the optimal schedule.
Evolutionary kind of algorithms is one of the best, but the time taken to findthe optimal schedule is more. This
paper presents a distance-based Pareto genetic algorithm(DPGA) with the Map Reduce model for
scheduling independent tasks in a DS environment. In DS, most of the task
scheduling problem is formulated as multi-objective optimization problem. This
paper aims to develop the optimal schedules by minimizing makespan and flow
time simultaneously. The algorithm is tested on a set of benchmark instances.
MapReduce model is used to parallelize the execution of DPGA automatically.
Experimental results show that DPGA with MapReduce model achieves a reduction
in makespan, mean flow time and execution time by 12%, 14% and 13% than
non-dominated sorting genetic algorithm (NSGA-II) with MapReduce model is also
implemented in this paper.