Creating a Schedule for Parallel Execution of Tasks Based on the Adjacency Lists

摘要

The article presents a method for transforming algorithm's information graph using adjacency lists. Algorithm's information graph always has a large number of vertices. For most algorithms, this graph contains more than 100 vertices. Manual analysis of this graph for the presence of internal parallelism is very difficult. The proposed method does not use conventional adjacency matrix for storing information about the connections between vertices and the adjacency lists. Adjacency lists allow to store information about the graph in a compressed form. As a result, the researcher gets a schedule of the algorithm on a computer, allowing parallel execution. The presented method can be successfully applied to queries in databases, to the distribution of tasks between nodes of a wireless network, to solving problems with large volumes of data in the field of the Internet of things.