An algorithm of JSSP with dynamic collection of job with priority

In this article a new algorithm for job-shop scheduling problem is proposed. In the algorithm a manufacturing tree is constructed on the basis of JSSP. Then the priorities of the nodes are set according to the levels. Except the condition that some operations being scheduled need dynamic adjusting, from the very beginning to the end one principle must be followed that is to keep the machine busy, which means that the job is dispatched to the machine incessantly so long as the machine is idle. During the scheduling an operation collection is dynamically generated according to the tree. Then some operation in the collection is scheduled according to the priority and other scheduling strategy (short-time strategy, long-path strategy, and dynamic-adjustment strategy) until the collection is empty, which means the entire job is finished. It is validated that the algorithm in the article is able to get better result for job-shop scheduling problem.