Two metaheuristic algorithms, TabuCol (Hertz et al., 1987) and simulated annealing (Johnson et al. 1991) , to find a good approximation of the chromatic number of two random graphs. The data here has the only goal of providing an example of use of eafplot for comparing algorithm performance with respect to both time and quality when modelled as two objectives in trade off.
Format
A data frame with 3133 observations on the following 6 variables.
alga factor with levels
SAKempeFIandTSinN1insta factor with levels
DSJC500.5andDSJC500.9. Instances are taken from the DIMACS repository.runa numeric vector indicating the run to which the observation belong.
besta numeric vector indicating the best solution in number of colors found in the corresponding run up to that time.
timea numeric vector indicating the time since the beginning of the run for each observation. A rescaling is applied.
titera numeric vector indicating iteration number corresponding to the observations.
Source
Marco Chiarandini (2005). Stochastic Local Search Methods for Highly Constrained Combinatorial Optimisation Problems. Ph.D. thesis, FB Informatik, TU Darmstadt, Germany. (page 138)
Details
Each algorithm was run 10 times per graph registering the time and iteration number at which a new best solution was found. A time limit corresponding to 500*10^5 total iterations of TabuCol was imposed. The time was then normalized on a scale from 0 to 1 to make it instance independent.
References
A. Hertz and D. de Werra. Using Tabu Search Techniques for Graph Coloring. Computing, 1987, 39(4), 345-351.
David S. Johnson, Cecilia R. Aragon, Lyle A. McGeoch, Catherine Schevon (1991). “Optimization by Simulated Annealing: An Experimental Evaluation: Part II, Graph Coloring and Number Partitioning.” Operations Research, 39(3), 378–406.
Examples
data(gcp2x2)