Automatic Graph Drawing and Stochastic Hill-Climbing

Alejandro Rosete-Suárez , Alberto Ochoa-Rodríguez , Michèle Sebag (1)

1 LMS - Laboratoire de mécanique des solides

In the literature of Evolutionary Computation, it is very strange to find papers where the results of Evolutionary Algorithms are compared to other algorithms. Stochastic Hill Climbing is a simple optimization algorithm that has shown a competitive performance with respect to many powerful algorithms in the solution of different problems. It has also outperformed some Evolutionary Algorithms in previous papers. Here we fairly review some of these papers. We also compare many Evolutionary Algorithms in the context of Graph Drawing. Graph Drawing addresses the problem of finding a representation of a graph that satisfies a given aesthetic objective. This problem has many practical applications in many fields such as Software Engineering, and VLSI Design. Our results demonstrate that Stochastic Hill Climbing is also the best algorithm in this context. We give some general guidelines in order to explain our results. Our explanations are based on landscape characteristics.

Domaines: Mécanique des solides [physics.class-ph]

Alejandro Rosete-Suárez, Alberto Ochoa-Rodríguez, Michèle Sebag. Automatic Graph Drawing and Stochastic Hill-Climbing. Genetic and Evolutionary Computation Conference, 1999, Paris, France. pp.1699-1706. ⟨hal-00116568⟩