Task Scheduling Model Based on Multi-Agent and Multi-Objective Dynamical Scheduling Algorithm

Abstract

With the increasing number of nodes in distributed systems, the complexity of task scheduling also increases. Therefore, how to schedule tasks reasonably is becoming more and more significant. Most traditional algorithms only consider a single condition without thinking over dynamic characteristic of system and tasks and lack of comprehensive measures. Therefore they cannot meet the needs of distributed systems. To solve these problems, we establish a distributed task scheduling model based on multi-agent in this paper, build a negotiation scheduling mechanism based on the model and propose distributed multi-objective dynamical scheduling algorithm (DMOD). In the algorithm, each node is capable of independent decision-making and dynamical evaluation rules make a comprehensive evaluation of task completion time, system load and communication traffic. DMOD, MinMin and the algorithm based on tree structure (BTS) are compared through simulation experiments. Experimental results show that DMOD reduces communication traffic without increasing task completion time, avoids performance degradation caused by sharp increase of system load and communication traffic in distributed system and therefore improves system stability and task execution efficiency.