| The Internet of things as a new kind of network architecture, widely used in various fields of national economy, in the process of the Internet of things applications, How together using the Internet of things application layer resources, to improve the information processing ability of the Internet of things application layer is more and more prominent problem, the key to solve this problem is the task scheduling, So the study of the task scheduling method based on Internet of things application layer to improve the efficiency of the operation of the Internet of things has great significance. This paper focuses on scheduling strategy, real-time and scheduling methods in the scheduling under the heterogeneous application environment of the Internet of things application layer for more in-depth analysis and research, the main work is as follows:1. Four kinds of the existing typical algorithms has been analyzed. the main idea of the four kinds of algorithms has been expounded, the complexity of these algorithms, the advantages of these algorithms, as well as disadvantages were analyzed. On this basis, the priority for these algorithms in the scheduling process did not show the dynamic changes. The target of using replication tasks method was not clear, did not reach the purpose of reducing scheduling length directly, not only increased the load resources, may also increased the final scheduling task scheduling etc, so this paper designed a new scheduling strategy.2. Based on this new strategy proposes a new scheduling algorithm IHDPTS (Improve Heterogeneous Dynamic priority task scheduling algorithm). The algorithm combined the advantages of list scheduling algorithm and the advantages of replication algorithm. through good design in advance scheduling priority, to schedule, and at the same time in the process of scheduling, using dynamic optimization strategy of task allocation priority,considering each resource load, as well as the tasks assigned to the resource impact on the subsequent tasks, and upward weights, task assignment priority was adjusted dynamically, rationalize, to ensure the task scheduling sequence during the scheduling. Starting from the export task node, to find key front task that restricts the early beginnings of the export task node, and then to find other key front task that restricts the the early beginnings of the key front node, At last, we find all key front tasks in this way. then according to these key front tasks, adopt reasonable and effective way of copy for further optimization of the earliest start of key front task execution time, so as to achieve the purpose of optimizing the whole task scheduling length.3. In view of the task scheduling algorithm and scheduling strategy that were proposed in this paper, this paper designed a test example that based on DAG, analyzed scheduling Process of test example. Through comparing and analyzing scheduling length and the change of the scheduling order in the scheduling process, this paper demonstrated that dynamic optimization scheduling sequence and finally by copying method optimization scheduling length is effective.4. A special task scheduling algorithm simulation software was developed, by generating a large number of different structures of the four groups of test cases, scheduling strategy and scheduling algorithm that were proposed in this paper, were confirmed, the results showed that this algorithm which was compared with existing scheduling algorithms had the characteristics of short the scheduling length, low complexity algorithm. |
