Graphs are structures used in different areas of scientific research, for the ease they have to represent different models of real life. There is a great variety of algorithms that build graphs with very dissimilar characteristics and types that model the relationships between the objects of the problem to solve. To model the relationships, characteristics such as depth, width and density of the graph are used in the directed acyclic graphs (DAGs) to find the solution to the objective problem.

more »

... hese characteristics are rarely analyzed and taken into account before being used in the approach of a solution. In this work, we present a set of methods for the random generation of DAGs. DAGs are produced with three of these methods representing three synthetic loads. Each of the three above characteristics is evaluated and analyzed in each of DAGs. The generation and evaluation of synthetic loads is with the objective of predicting the behavior of each DAG, based on its characteristics, in a scheduling algorithm and assignment of parallel tasks in a distributed heterogeneous computing system (DHCS). test DAGs that are received in the target system to measure different performance metrics.