A New Heuristic Layout Algorithm for DAGs

Directed acyclic graphs (DAGs) are a common representation of hierarchical relationships and are widely used in software and information engineering, system theory, and others. The usefulness depends on a readable, understandable and easy to remember layout. The key problem in making such a good layout is the reduction of numbers of arc crossings. We focus on the two layer problem and make a fresh attempt at a heuristic method of arc crossing reduction based on a stochastic approach. The “stochastic” heuristic is described in detail. The results of the approach are analyzed. A comparison of these results with the mostly used methods of barycentering and median-ordering is given. The “stochastic” heuristic is shown to be preferable in terms of Warfield’s generating matrices.