Exploratory Analysis of Stochastic Local Search Algorithms in Biobjective Optimization

This chapter introduces two Perl programs that implement graphical tools for exploring the performance of stochastic local search algorithms for biobjective optimization problems. These tools are based on the concept of the empirical attainment function (EAF), which describes the probabilistic distribution of the outcomes obtained by a stochastic algorithm in the objective space. In particular, we consider the visualization of attainment surfaces and differences between the first-order EAFs of the outcomes of two algorithms. This visualization allows us to identify certain algorithmic behaviors in a graphical way. We explain the use of these visualization tools and illustrate them with examples arising from practice. Experiments in computer science often produce large amounts of data, mainly because experiments can be set up, performed and repeated with relative facility. Given the amount of data, exploratory data analysis techniques are one of the most important tools that computer scientists may use to support their findings. In particular, specialized graphical techniques for representing data are often used to perceive trends and patterns in the data. For instance, there exist techniques for the extraction of relevant variables, the discovery of hidden structures, and the detection of outliers and other anomalies. Such exploratory techniques are mainly used during the design of an algorithm and when comparing the performance of various algorithms. Even before testing more formal hypotheses, the algorithm designer has to find patterns in experimental data that provide further insights into new ways of improving performance. In this chapter, we focus on the graphical interpretation of the quality of the outcomes returned by Stochastic Local Search (SLS) algorithms [Hoos and Stützle, Luı́s Paquete CISUC, Department of Informatics Engineering, University of Coimbra, Portugal e-mail: [email protected] Manuel López-Ibáñez, Thomas Stützle IRIDIA, CoDE, Université Libre de Bruxelles, Brussels, Belgium e-mail: [email protected],[email protected]